DRSSTC Peak Primary Current formula?
Sigurthr, Thu Jan 23 2014, 12:46AMHey everyone!
I'm considering making my "Big 3kW SSTC" into a DR coil since I'm rebuilding the primary side and driver after a near-fatal flashover when the old primary overheated last month.
Is there a formula for determining Peak Primary variables? I'd like to figure out how long of a pulse width I can use to stay under a certain peak primary current, and if possible, determine what the peak primary voltage will be so I can properly design the MMC.
Here's data on my coil:
J A V A T C version 13.2 - CONSOLIDATED OUTPUT
Wednesday, January 22, 2014 6:22:30 PM
Units = Inches
Ambient Temp = 68°F
------------------------------------------- ---------
Secondary Coil Inputs:
------------------------------------------ ----------
Current Profile = G.PROFILE_LOADED
2.25 = Radius 1
2.25 = Radius 2
3.5 = Height 1
21.25 = Height 2
1505 = Turns
30 = Wire Awg
--------------------------------------------- -------
Primary Coil Inputs:
------------------------------------------ ----------
Round Primary Conductor
3.125 = Radius 1
3.125 = Radius 2
1 = Height 1
2.8 = Height 2
10 = Turns
12 = Wire Awg
0 = Ribbon Width
0 = Ribbon Thickness
0.047 = Primary Cap (uF)
30 = Total Lead Length
12 = Lead Diameter
---------------------------------------- ------------
Top Load Inputs:
------------------------------------------ ----------
Toroid #1: minor=3, major=12, height=24.75, topload
----------------------------------------- -----------
Secondary Outputs:
----------------------------------------- -----------
172.12 kHz = Secondary Resonant Frequency
90 deg° = Angle of Secondary
17.75 inch = Length of Winding
84.8 inch = Turns Per Unit
0.00177 inch = Space Between Turns (edge to edge)
1773 ft = Length of Wire
3.94:1 = H/D Aspect Ratio
181.4684 Ohms = DC Resistance
63705 Ohms = Reactance at Resonance
0.54 lbs = Weight of Wire
58.906 mH = Les-Effective Series Inductance
60.636 mH = Lee-Equivalent Energy Inductance
58.794 mH = Ldc-Low Frequency Inductance
14.515 pF = Ces-Effective Shunt Capacitance
14.101 pF = Cee-Equivalent Energy Capacitance
23.583 pF = Cdc-Low Frequency Capacitance
7.23 mils = Skin Depth
10.1 pF = Topload Effective Capacitance
261.8198 Ohms = Effective AC Resistance
243 = Q
----------------------------------------------- -----
Primary Outputs:
----------------------------------------- -----------
157.6 kHz = Primary Resonant Frequency
8.44 % high = Percent Detuned
90 deg° = Angle of Primary
16.36 ft = Length of Wire
25.99 mOhms = DC Resistance
0.099 inch = Average spacing between turns (edge to edge)
1.075 inch = Proximity between coils
0 inch = Recommended minimum proximity between coils
21.434 µH = Ldc-Low Frequency Inductance
0.0394 µF = Cap size needed with Primary L (reference)
0.266 µH = Lead Length Inductance
104.31 µH = Lm-Mutual Inductance
0.093 k = Coupling Coefficient
0.131 k = Recommended Coupling Coefficient
10.75 = Number of half cycles for energy transfer at K
33.93 µs = Time for total energy transfer (ideal quench time)
I messed around with the JavaDRC calculator, but I'm unfamiliar with it, and it doesn't seem to take in account the bus voltage, half or full bridge inverter, etc, which all seem to be rather important for determining peak currents and voltages. Here are the outputs anyway.
J A V A D R C - CONSOLIDATED OUTPUT
Wednesday, January 22, 2014 6:47:40 PM
Data Inputs:
0.047 [uF] = Single Cap Capacitance value
1 [qty] = Number of Caps in Series String
1 [qty] = Number of Strings in MMC
1414 [Vac] = Single Capacitors AC Voltage Rating
22 [uH] = Primary Inductance
170000 [Hz] = Secondary Resonant Frequency
90 [Amps] = Expected Peak Current
350 [uS] = Silicon Pulse On Time
100 [pps] = Silicon Pulses per Second
------------------------------------------ ----------
Data Outputs:
0.047 [uF] = Total Cap Bank Capacitance
2000 [Vp] = Cap String Rated Peak Voltage
1947 [Vp] = Expected Peak Voltage
97 [%] = Percentage of Cap String Rating to Expected Peak Voltage
0.09 [joules] = Peak Energy at Expected Current
21.64 [ohms] = Tank Surge Impedance
156516 [Hz] = Primary Resonant Frequency
170000 [Hz] = Secondary Resonant Frequency
8 [%] = Primary to Secondary Percent Detuned
14 [Amps] = RMS Current Per Silicon On Time and PPS
I'll be using a Half-Bridge fed by full wave rectified and filtered 120V mains with variac control. My interrupter will be set for 100Hz and I'll have precision variable ON pulse width. I'll be using FDL100N50F MOSFETs so I want to stay under 100Apk. I know my Fres will be 170KHz, so if I can figure out the current and voltage rise per cycle of ON TIME then I can determine the ON TIME limit to set. For example, at 5.88uS/cycle then 350uS ON will yield 60 cycles. I just need to ballpark how many cycles (how long of ON TIME) to set the interrupter for, for the first light. I'd hate to blow $30 of silicon in the first mS of run time because I was off by a factor of ten by arbitrarily picking a value.
I'll be happy if I can achieve 6" breakout as that's how much spark length I got running this coil ISSTCC at 1.5kW.
Re: DRSSTC Peak Primary Current formula?
Mads Barnkob, Thu Jan 23 2014, 06:02AM
That is pretty much what I made the MMC calculator for:
Mads Barnkob, Thu Jan 23 2014, 06:02AM
That is pretty much what I made the MMC calculator for:
Re: DRSSTC Peak Primary Current formula?
Sigurthr, Thu Jan 23 2014, 07:40AM
Forgive my confusion, but it seems that calc assumes you know your peak current. It is nice that it gives you the peak voltage across each cap and dV/dT though! How do you determine how much your primary current rings up per cycle in order to determine what the peak current will be for a given on time?
Sigurthr, Thu Jan 23 2014, 07:40AM
Forgive my confusion, but it seems that calc assumes you know your peak current. It is nice that it gives you the peak voltage across each cap and dV/dT though! How do you determine how much your primary current rings up per cycle in order to determine what the peak current will be for a given on time?
Re: DRSSTC Peak Primary Current formula?
Uspring, Thu Jan 23 2014, 10:26AM
Basically you'll add the switching voltage to your MMC voltage every time you switch. In detail:
With a mains voltage of 120V and a half bridge you switch between -170V and +170V. Initially you start off with 0V. When voltage is first applied, the MMC voltage will rise to 170V after one half cycle, i.e. 3us. Then you switch input to -170V. After 3us MMC voltage will be 170V + 340V. You have added the switching amplitude (340V) to the initial voltage. After another 3us MMC voltage wil be 170V + 2*340V and so on.
The MMC voltage determines the current:
I = V / (2* Pi * Lpri * fres)
with primary fres. For 21uH and 160kHz, 100A will be reached, when the MMC voltage is 2100V, i.e. after about 6 half cycles or about 18us.
This is very much an upper limit for the current. Losses in the primary tank and even more the loading due to the secondary will limit the current. If you have a scope, I'd monitor it and go from there.
I believe these fets can stand much more current for the short burst times you are aiming at.
Uspring, Thu Jan 23 2014, 10:26AM
Basically you'll add the switching voltage to your MMC voltage every time you switch. In detail:
With a mains voltage of 120V and a half bridge you switch between -170V and +170V. Initially you start off with 0V. When voltage is first applied, the MMC voltage will rise to 170V after one half cycle, i.e. 3us. Then you switch input to -170V. After 3us MMC voltage will be 170V + 340V. You have added the switching amplitude (340V) to the initial voltage. After another 3us MMC voltage wil be 170V + 2*340V and so on.
The MMC voltage determines the current:
I = V / (2* Pi * Lpri * fres)
with primary fres. For 21uH and 160kHz, 100A will be reached, when the MMC voltage is 2100V, i.e. after about 6 half cycles or about 18us.
This is very much an upper limit for the current. Losses in the primary tank and even more the loading due to the secondary will limit the current. If you have a scope, I'd monitor it and go from there.
I believe these fets can stand much more current for the short burst times you are aiming at.
Re: DRSSTC Peak Primary Current formula?
Dr. Dark Current, Thu Jan 23 2014, 08:01PM
For (Q)CW coils I have always estimated the primary current by the Q of the primary tank circuit. For the CW coils the lowest vaule seems to be around 5, I usually calculate for 10 and then detune to get the required current. For a pulsed DR, the Q might reach around 10-15(?) during the first peak.
Dr. Dark Current, Thu Jan 23 2014, 08:01PM
For (Q)CW coils I have always estimated the primary current by the Q of the primary tank circuit. For the CW coils the lowest vaule seems to be around 5, I usually calculate for 10 and then detune to get the required current. For a pulsed DR, the Q might reach around 10-15(?) during the first peak.
Re: DRSSTC Peak Primary Current formula?
Mads Barnkob, Thu Jan 23 2014, 08:12PM
I read your thread on the phone at breakfast, missed half of the question in my eager to post a answer before rushing out the door for work :)
Mads Barnkob, Thu Jan 23 2014, 08:12PM
Sigurthr wrote ...
Forgive my confusion, but it seems that calc assumes you know your peak current. It is nice that it gives you the peak voltage across each cap and dV/dT though! How do you determine how much your primary current rings up per cycle in order to determine what the peak current will be for a given on time?
Forgive my confusion, but it seems that calc assumes you know your peak current. It is nice that it gives you the peak voltage across each cap and dV/dT though! How do you determine how much your primary current rings up per cycle in order to determine what the peak current will be for a given on time?
I read your thread on the phone at breakfast, missed half of the question in my eager to post a answer before rushing out the door for work :)
Re: DRSSTC Peak Primary Current formula?
Sigurthr, Thu Jan 23 2014, 08:24PM
Ahh! Thank you guys very much!!!
Yep I plan on scoping it out but I don't have a storage oscilloscope, just an old analog CRT one, so I typically can't see pulsed or transient effects that happen less frequently than 200Hz or so.
I'm going to have to enlarge my MMC if voltage rings up that fast. These fets can handle either 400 or 500amps pulsed but my mmc was only rated for 2kV! 18uS sounds like an incredibly short burst length, I usually hear of people running hundreds of uS.
Sigurthr, Thu Jan 23 2014, 08:24PM
Ahh! Thank you guys very much!!!
Yep I plan on scoping it out but I don't have a storage oscilloscope, just an old analog CRT one, so I typically can't see pulsed or transient effects that happen less frequently than 200Hz or so.
I'm going to have to enlarge my MMC if voltage rings up that fast. These fets can handle either 400 or 500amps pulsed but my mmc was only rated for 2kV! 18uS sounds like an incredibly short burst length, I usually hear of people running hundreds of uS.
Re: DRSSTC Peak Primary Current formula?
Uspring, Fri Jan 24 2014, 09:34AM
Also without an OCD this is hazardous, since a ground arc throws the secondary out of tune and will allow almost unlimited current rise.
Uspring, Fri Jan 24 2014, 09:34AM
18uS sounds like an incredibly short burst length, I usually hear of people running hundreds of uS.A 170V, 100A half bridge can supply more than 10kW of power, while a 2kV 47nF MMC only stores about 0.1J. It is not surprising, that you can run only short bursts unless the secondary draws a significant part of the power. It might be possible to tune the coil, so that 100A won't be exceeded in longer burst. But I have doubts about this, since the arc would have to break out very fast, i.e. faster than 18us.
Also without an OCD this is hazardous, since a ground arc throws the secondary out of tune and will allow almost unlimited current rise.
Re: DRSSTC Peak Primary Current formula?
Sigurthr, Fri Jan 24 2014, 10:51PM
Yeah. I was hoping for a quick and easy conversion to DR; add MMC and special interrupter to limit burst length. Unless I can reliably limit burst length to ~15uS (a tall order in of itself) and make sure ouput stays under 20" I'm going to need a complex controller with OCD and synchronized turn off (to prevent a hard switch), and I have no idea how to implement such a synchronized stop, haha. Hell, I'm not sure I can get my slow MCU to do accurate 15uS pulses.
Sigurthr, Fri Jan 24 2014, 10:51PM
Yeah. I was hoping for a quick and easy conversion to DR; add MMC and special interrupter to limit burst length. Unless I can reliably limit burst length to ~15uS (a tall order in of itself) and make sure ouput stays under 20" I'm going to need a complex controller with OCD and synchronized turn off (to prevent a hard switch), and I have no idea how to implement such a synchronized stop, haha. Hell, I'm not sure I can get my slow MCU to do accurate 15uS pulses.
Re: DRSSTC Peak Primary Current formula?
Sigurthr, Wed Jan 29 2014, 01:30AM
Bear with me if you would be so kind please, I want to verify the timeline here so I know I have a good grasp of it. I'm not quite following the wording of the mathematical function.
Based on "adding switching voltage to MMC voltage every time you switch" I originally thought:
1) Voltage first applied; MMC voltage: 0V. Time Since Start: 0uS
2) First half cycle complete. MMC voltage: 170V. Time Since Start: 3.125uS
3) Second half cycle complete. MMC voltage: 340V. Time Since Start: 6.25uS.
4) 3rd half cycle complete. MMC voltage: 510V. Time Since Start: 9.375uS.
5) 4th half cycle complete. MMC voltage: 680V. Time Since Start: 12.5uS
With each new line being a half cycle:
170
170 170
170 170 170
170 170 170 170
With the timeline I spelled out based on beginning of your description it seems like each cycle 2x bus voltage is added to whatever the MMC was at prior. However, at the middle and end of your description it seems there is an offset to my timeline; You say that at 3.125uS the voltage will be at 170V, then at 6.25uS the voltage will be 510V, then at the 9.375uS mark the voltage will be 850V.
This looks like:
170
170 170 170
170 170 170 170 170
Based on that pattern we get:
170
170 170 170
170 170 170 170 170
170 170 170 170 170 170 170
170 170 170 170 170 170 170 170 170
170 170 170 170 170 170 170 170 170 170 170
This shows that each half cycle after the first we are adding 2x the bus voltage to what the MMC voltage was the previous half cycle. Is this right? If so then we can extrapolate:
V= Vin x 1 = Vin x (N + 0)
V= Vin x 3 = Vin x (N + 1)
V= Vin x 5 = Vin x (N + 2)
V= Vin x 7 = Vin x (N + 3)
V= Vin x 9 = Vin x (N + 4)
V= Vin x 11 = Vin x (N + 5)
which yields the following formula for MMC voltage at a given On Time at a given Input Voltage:
V = Vin x (N + (N-1))
where N = number of half cycles
Math is regrettably my weakest foundation, so if this looks horribly wrong I do apologize and humbly ask that the correction be spelled out in painfully clear detail. My sincerest gratitude for your time, patience, and help.
Sigurthr, Wed Jan 29 2014, 01:30AM
Uspring wrote ...
Basically you'll add the switching voltage to your MMC voltage every time you switch. In detail:
With a mains voltage of 120V and a half bridge you switch between -170V and +170V. Initially you start off with 0V. When voltage is first applied, the MMC voltage will rise to 170V after one half cycle, i.e. 3us. Then you switch input to -170V. After 3us MMC voltage will be 170V + 340V. You have added the switching amplitude (340V) to the initial voltage. After another 3us MMC voltage wil be 170V + 2*340V and so on.
Basically you'll add the switching voltage to your MMC voltage every time you switch. In detail:
With a mains voltage of 120V and a half bridge you switch between -170V and +170V. Initially you start off with 0V. When voltage is first applied, the MMC voltage will rise to 170V after one half cycle, i.e. 3us. Then you switch input to -170V. After 3us MMC voltage will be 170V + 340V. You have added the switching amplitude (340V) to the initial voltage. After another 3us MMC voltage wil be 170V + 2*340V and so on.
Bear with me if you would be so kind please, I want to verify the timeline here so I know I have a good grasp of it. I'm not quite following the wording of the mathematical function.
Based on "adding switching voltage to MMC voltage every time you switch" I originally thought:
1) Voltage first applied; MMC voltage: 0V. Time Since Start: 0uS
2) First half cycle complete. MMC voltage: 170V. Time Since Start: 3.125uS
3) Second half cycle complete. MMC voltage: 340V. Time Since Start: 6.25uS.
4) 3rd half cycle complete. MMC voltage: 510V. Time Since Start: 9.375uS.
5) 4th half cycle complete. MMC voltage: 680V. Time Since Start: 12.5uS
With each new line being a half cycle:
170
170 170
170 170 170
170 170 170 170
With the timeline I spelled out based on beginning of your description it seems like each cycle 2x bus voltage is added to whatever the MMC was at prior. However, at the middle and end of your description it seems there is an offset to my timeline; You say that at 3.125uS the voltage will be at 170V, then at 6.25uS the voltage will be 510V, then at the 9.375uS mark the voltage will be 850V.
This looks like:
170
170 170 170
170 170 170 170 170
Based on that pattern we get:
170
170 170 170
170 170 170 170 170
170 170 170 170 170 170 170
170 170 170 170 170 170 170 170 170
170 170 170 170 170 170 170 170 170 170 170
This shows that each half cycle after the first we are adding 2x the bus voltage to what the MMC voltage was the previous half cycle. Is this right? If so then we can extrapolate:
V= Vin x 1 = Vin x (N + 0)
V= Vin x 3 = Vin x (N + 1)
V= Vin x 5 = Vin x (N + 2)
V= Vin x 7 = Vin x (N + 3)
V= Vin x 9 = Vin x (N + 4)
V= Vin x 11 = Vin x (N + 5)
which yields the following formula for MMC voltage at a given On Time at a given Input Voltage:
V = Vin x (N + (N-1))
where N = number of half cycles
Math is regrettably my weakest foundation, so if this looks horribly wrong I do apologize and humbly ask that the correction be spelled out in painfully clear detail. My sincerest gratitude for your time, patience, and help.
Re: DRSSTC Peak Primary Current formula?
Uspring, Wed Jan 29 2014, 10:59AM
0 170 510 850 1190 ...
This is only true if all the bridge power goes solely into the primary tank and the primary is lossless. Coupling to the secondary will reduce the voltage and primary current. That depends on the amount of coupling and tuning.
Uspring, Wed Jan 29 2014, 10:59AM
V = Vin x (N + (N-1))This is correct except for N=0. The change in the bridges output voltage is added every half cycle. The first half cycle is special, since the voltage change is from 0 to 170. For the next half cycle bridge voltage jumps from 170 to -170 so the change in bridge output voltage is 340. This leads to MMC voltages in the pattern:
where N = number of half cycles
0 170 510 850 1190 ...
This is only true if all the bridge power goes solely into the primary tank and the primary is lossless. Coupling to the secondary will reduce the voltage and primary current. That depends on the amount of coupling and tuning.
Re: DRSSTC Peak Primary Current formula?
Sigurthr, Wed Jan 29 2014, 07:25PM
Excellent! Right, it doesn't work for N=0, but I'm not overly concerned about that. My goal is to make a calculator that will determine the maximum burst length a MMC can withstand without being over-volted using worst case conditions (lossless primary, no power transfer, etc). The reason behind it is for those of us looking to build a DRSSTC with a very small MMC due to budget it is important to know a ball-park value for how long of a burst is safe to run. If I hadn't come to 4HV and asked about this I'd have fired up my DRSSTC for a 50uS burst the first time when a 18uS burst is enough to over-volt the MMC!
I'm surprised a half-bridge adds 2x Vbus per half cycle when in a standard nonresonant environment the load only sees 1/2 Vbus as Vpk (or Vbus as Vp-p).
Sigurthr, Wed Jan 29 2014, 07:25PM
Excellent! Right, it doesn't work for N=0, but I'm not overly concerned about that. My goal is to make a calculator that will determine the maximum burst length a MMC can withstand without being over-volted using worst case conditions (lossless primary, no power transfer, etc). The reason behind it is for those of us looking to build a DRSSTC with a very small MMC due to budget it is important to know a ball-park value for how long of a burst is safe to run. If I hadn't come to 4HV and asked about this I'd have fired up my DRSSTC for a 50uS burst the first time when a 18uS burst is enough to over-volt the MMC!
I'm surprised a half-bridge adds 2x Vbus per half cycle when in a standard nonresonant environment the load only sees 1/2 Vbus as Vpk (or Vbus as Vp-p).
Re: DRSSTC Peak Primary Current formula?
Uspring, Thu Jan 30 2014, 09:09AM
A non-resonant analogon would be an inductance charged up with a Vin DC voltage. Current would rise according to
I = V * t / L
100A are reached then in about 12us with 170V and 21uH. That's a bit shorter than the 18us for the resonant case. This is due to the intermediate drops of primary current to zero in the resonant case. When current is low there is little power transfer from the voltage source to the tank.
Uspring, Thu Jan 30 2014, 09:09AM
I'm surprised a half-bridge adds 2x Vbus per half cycle when in a standard nonresonant environment the load only sees 1/2 Vbus as Vpk (or Vbus as Vp-p).I'm confused. Isn't Vp-p equal to 2*Vbus? I'm supposing that Vbus is the same as Vin.
A non-resonant analogon would be an inductance charged up with a Vin DC voltage. Current would rise according to
I = V * t / L
100A are reached then in about 12us with 170V and 21uH. That's a bit shorter than the 18us for the resonant case. This is due to the intermediate drops of primary current to zero in the resonant case. When current is low there is little power transfer from the voltage source to the tank.
Re: DRSSTC Peak Primary Current formula?
Sigurthr, Thu Jan 30 2014, 10:37AM
Well what I meant was that in a standard SSTC (or any nonresonant load) on a half bridge the primary/load sees alternating +85V and -85V, which is 170Vp-p, which is 1x the total bus voltage.
In the DRSSTC On a half bridge model described above after the first half cycle the LC circuit sees an increase of 340V per half cycle over the previous half cycle. This is 2x Vbus.
I think for a full bridge on rectified 120V mains the load sees Vp-p = 2Vbus = 340V.
So what confused me was that Vp-p across the load seems to have magically doubled by going from nonresonant to resonant load conditions.
Sigurthr, Thu Jan 30 2014, 10:37AM
Well what I meant was that in a standard SSTC (or any nonresonant load) on a half bridge the primary/load sees alternating +85V and -85V, which is 170Vp-p, which is 1x the total bus voltage.
In the DRSSTC On a half bridge model described above after the first half cycle the LC circuit sees an increase of 340V per half cycle over the previous half cycle. This is 2x Vbus.
I think for a full bridge on rectified 120V mains the load sees Vp-p = 2Vbus = 340V.
So what confused me was that Vp-p across the load seems to have magically doubled by going from nonresonant to resonant load conditions.
Re: DRSSTC Peak Primary Current formula?
Uspring, Thu Jan 30 2014, 02:02PM
I think I have misunderstood you all along. You wrote:
Sorry about this.
Uspring, Thu Jan 30 2014, 02:02PM
I think I have misunderstood you all along. You wrote:
I'll be using a Half-Bridge fed by full wave rectified and filtered 120V mainsI was thinking your half bridge was switching between 2 (half wave rectified) rails of +170V and -170V. For full wave rectified 170V, switching takes place between 0 and 170V and your voltage rampup is only at half the speed, i.e. 170V per half cycle.
Sorry about this.
Re: DRSSTC Peak Primary Current formula?
Sigurthr, Thu Jan 30 2014, 11:54PM
No worries! You've still provided immense help this whole time.
So then, I should be able to amend the formula for a half bridge to:
Vmmc = 1/2 Vbus * ((2N)-1)
and then for a full bridge it would be:
Vmmc = Vbus * ((2N)-1)
This actually works out to my benefit as I can easily add an option to select whether a half or full bridge is used in the calculator I'm making. Also, I can have twice as long of a burst length for a half bridge.
The other formula shouldn't need any modification, right? It makes sense that it wouldn't as it describes the voltage at a certain peak current for the resonant circuit, and isn't related to the input voltage.
V = Ipk * 2Pi * Lpri * Fres
Sigurthr, Thu Jan 30 2014, 11:54PM
Uspring wrote ...
I think I have misunderstood you all along. You wrote:
Sorry about this.
I think I have misunderstood you all along. You wrote:
I'll be using a Half-Bridge fed by full wave rectified and filtered 120V mainsI was thinking your half bridge was switching between 2 (half wave rectified) rails of +170V and -170V. For full wave rectified 170V, switching takes place between 0 and 170V and your voltage rampup is only at half the speed, i.e. 170V per half cycle.
Sorry about this.
No worries! You've still provided immense help this whole time.
So then, I should be able to amend the formula for a half bridge to:
Vmmc = 1/2 Vbus * ((2N)-1)
and then for a full bridge it would be:
Vmmc = Vbus * ((2N)-1)
This actually works out to my benefit as I can easily add an option to select whether a half or full bridge is used in the calculator I'm making. Also, I can have twice as long of a burst length for a half bridge.
The other formula shouldn't need any modification, right? It makes sense that it wouldn't as it describes the voltage at a certain peak current for the resonant circuit, and isn't related to the input voltage.
V = Ipk * 2Pi * Lpri * Fres
Re: DRSSTC Peak Primary Current formula?
Uspring, Fri Jan 31 2014, 02:07PM
Uspring, Fri Jan 31 2014, 02:07PM
So then, I should be able to amend the formula for a half bridge to:Almost. Since the voltage swing is Vbus including the first half cycle, it's simply Vbus * N and for the full bridge Vbus * 2N.
Vmmc = 1/2 Vbus * ((2N)-1)
The other formula shouldn't need any modification, right? It makes sense that it wouldn't as it describes the voltage at a certain peak current for the resonant circuit, and isn't related to the input voltage.Right.
V = Ipk * 2Pi * Lpri * Fres
Re: DRSSTC Peak Primary Current formula?
Kizmo, Fri Jan 31 2014, 04:14PM
Interesting thing is that the primary current can be larger than this math suggests too. All depends on your load conditions and tuning point. I think under some conditions the secondary can back drive the primary resulting higher currents.
Kizmo, Fri Jan 31 2014, 04:14PM
Interesting thing is that the primary current can be larger than this math suggests too. All depends on your load conditions and tuning point. I think under some conditions the secondary can back drive the primary resulting higher currents.
Re: DRSSTC Peak Primary Current formula?
Dr. Dark Current, Fri Jan 31 2014, 10:18PM
I think there is no way for it to be larger, only smaller, as spark loading drops the Q of the tank circuit.
Dr. Dark Current, Fri Jan 31 2014, 10:18PM
I think there is no way for it to be larger, only smaller, as spark loading drops the Q of the tank circuit.
Re: DRSSTC Peak Primary Current formula?
Sigurthr, Fri Jan 31 2014, 11:02PM
Are you sure? This is basically describing what I initially thought would happen until you stated otherwise:
If that's the case it certainly simplifies things. I wish I could simply test a resonant half bridge and verify these suppositions but I'm just not equipped to do so. I have a software engineer friend doing the programming for this calculator I want to make and I'm sure I'm driving him batty with the constant revisions, haha. Though, an issue we came across in implementation was determining a good way of finding the largest N (Nmax) before the voltage across the Primary Capacitor (Vmmc) exceeds the rated voltage (VmmcMAX). The only way we came up with was setting a bound of 20,000 for N (10,000 cycles) and using a Loop that counts down from 20k until Vmmc < VmmcMAX returns true. While this brute force method works it isn't the simplest and certainly isn't efficient.
If it really is just Vmmc = Vbus * N for half-bridge then we can simply just take the integer of ((VmmcMAX / Vbus) -1) to get a conservative Nmax. I believe the code to truncate from the float N to the integer Nmax would simply be: Nmax = (int) N; and this is a whole lot faster and easier than a complicated brute force search for Nmax.
Re: larger primary currents;
The only time I know of where primary currents may exceed the predicted when loading is factored is in when the drive frequency shifts away from the secondary resonance frequency resulting in lessened loading. For example if you were using secondary base feedback and had the primary tuned to exact resonance - as the Fres of the secondary drops from streamer capacitance it would take more primary current to transfer the same energy to the secondary. This is one reason why primary feedback and a lower tuned primary is often used, to compensate for this and to transfer the most energy into the arc once breakout has been achieved.
Sigurthr, Fri Jan 31 2014, 11:02PM
Uspring wrote ...
So then, I should be able to amend the formula for a half bridge to:Almost. Since the voltage swing is Vbus including the first half cycle, it's simply Vbus * N and for the full bridge Vbus * 2N.
Vmmc = 1/2 Vbus * ((2N)-1)
Are you sure? This is basically describing what I initially thought would happen until you stated otherwise:
Sigurthr wrote ...
Based on "adding switching voltage to MMC voltage every time you switch" I originally thought:
1) Voltage first applied; MMC voltage: 0V. Time Since Start: 0uS
2) First half cycle complete. MMC voltage: 170V. Time Since Start: 3.125uS
3) Second half cycle complete. MMC voltage: 340V. Time Since Start: 6.25uS.
4) 3rd half cycle complete. MMC voltage: 510V. Time Since Start: 9.375uS.
5) 4th half cycle complete. MMC voltage: 680V. Time Since Start: 12.5uS
With each new line being a half cycle:
170
170 170
170 170 170
170 170 170 170
Based on "adding switching voltage to MMC voltage every time you switch" I originally thought:
1) Voltage first applied; MMC voltage: 0V. Time Since Start: 0uS
2) First half cycle complete. MMC voltage: 170V. Time Since Start: 3.125uS
3) Second half cycle complete. MMC voltage: 340V. Time Since Start: 6.25uS.
4) 3rd half cycle complete. MMC voltage: 510V. Time Since Start: 9.375uS.
5) 4th half cycle complete. MMC voltage: 680V. Time Since Start: 12.5uS
With each new line being a half cycle:
170
170 170
170 170 170
170 170 170 170
If that's the case it certainly simplifies things. I wish I could simply test a resonant half bridge and verify these suppositions but I'm just not equipped to do so. I have a software engineer friend doing the programming for this calculator I want to make and I'm sure I'm driving him batty with the constant revisions, haha. Though, an issue we came across in implementation was determining a good way of finding the largest N (Nmax) before the voltage across the Primary Capacitor (Vmmc) exceeds the rated voltage (VmmcMAX). The only way we came up with was setting a bound of 20,000 for N (10,000 cycles) and using a Loop that counts down from 20k until Vmmc < VmmcMAX returns true. While this brute force method works it isn't the simplest and certainly isn't efficient.
If it really is just Vmmc = Vbus * N for half-bridge then we can simply just take the integer of ((VmmcMAX / Vbus) -1) to get a conservative Nmax. I believe the code to truncate from the float N to the integer Nmax would simply be: Nmax = (int) N; and this is a whole lot faster and easier than a complicated brute force search for Nmax.
Re: larger primary currents;
The only time I know of where primary currents may exceed the predicted when loading is factored is in when the drive frequency shifts away from the secondary resonance frequency resulting in lessened loading. For example if you were using secondary base feedback and had the primary tuned to exact resonance - as the Fres of the secondary drops from streamer capacitance it would take more primary current to transfer the same energy to the secondary. This is one reason why primary feedback and a lower tuned primary is often used, to compensate for this and to transfer the most energy into the arc once breakout has been achieved.
Re: DRSSTC Peak Primary Current formula?
Uspring, Sat Feb 01 2014, 07:14PM
Kizmo wrote:
Uspring, Sat Feb 01 2014, 07:14PM
Are you sure? This is basically describing what I initially thought would happen until you stated otherwise:Reasonably so. The "corrections" I made were due to the wrong assumption of switching between +170V and -170V rails.
Kizmo wrote:
Interesting thing is that the primary current can be larger than this math suggests too. All depends on your load conditions and tuning point. I think under some conditions the secondary can back drive the primary resulting higher currents.The only condition I can think up for this is a short cut secondary. That would not cause any losses but reduce primary inductance, i.e. higher primary frequency. Cycles are then shorter and rampup is faster. It is a small effect since coupling is low. A sensibly tuned secondary which isn't short cut avoids this.
Re: DRSSTC Peak Primary Current formula?
Kizmo, Sat Feb 01 2014, 07:43PM
Something like this happened when i first ran my big coil. More than likely the tuning was way off but with secondary in place i was able to trip the over current protection which did not happen when i tested it without secondary.
Who knows, might have been something else too.. :)
Kizmo, Sat Feb 01 2014, 07:43PM
Something like this happened when i first ran my big coil. More than likely the tuning was way off but with secondary in place i was able to trip the over current protection which did not happen when i tested it without secondary.
Who knows, might have been something else too.. :)
Re: DRSSTC Peak Primary Current formula?
Dr. Dark Current, Sat Feb 01 2014, 11:12PM
If you stored more energy in the primary tank cap than you put in, well that would be, say it .. a perpetuum mobile
Dr. Dark Current, Sat Feb 01 2014, 11:12PM
If you stored more energy in the primary tank cap than you put in, well that would be, say it .. a perpetuum mobile
Re: DRSSTC Peak Primary Current formula?
Sigurthr, Wed Jul 23 2014, 06:37AM
Hey everyone, hope you don't mind the slightly necropost. I do have new content to add though, I'm quite happy to announce!
I've been studying C++ programming lately and have learned enough to write a basic calculator program for the formulae discussed in this thread. I've tested it on all my PCs, which are windows 7 and 8.1, so I cannot guarantee that it will work on any other platform or OS. Please let me know what you think.
Here's the link to download the .rar
EDIT: Fixed the link, now points to the correct release file! Sorry folks!
Just for reference; here are the formulae I used:
Ipk = V / (2Pi * Lpri * fres)
N = Vlimit / Vbus
(half bridge)
N = (Vlimit / Vbus) / 2
(full bridge)
Where:
N = number of switched half cycles. Vlimit = chosen peak tank voltage.
Sigurthr, Wed Jul 23 2014, 06:37AM
Hey everyone, hope you don't mind the slightly necropost. I do have new content to add though, I'm quite happy to announce!
I've been studying C++ programming lately and have learned enough to write a basic calculator program for the formulae discussed in this thread. I've tested it on all my PCs, which are windows 7 and 8.1, so I cannot guarantee that it will work on any other platform or OS. Please let me know what you think.
Here's the link to download the .rar
EDIT: Fixed the link, now points to the correct release file! Sorry folks!
Just for reference; here are the formulae I used:
Ipk = V / (2Pi * Lpri * fres)
N = Vlimit / Vbus
(half bridge)
N = (Vlimit / Vbus) / 2
(full bridge)
Where:
N = number of switched half cycles. Vlimit = chosen peak tank voltage.
Re: DRSSTC Peak Primary Current formula?
DerStrom8, Wed Jul 23 2014, 12:55PM
Hi Sigurthr, thanks for posting.
I get an error when I try to run the exe on Windows 8:
I did some quick research and found this:
Hopefully this will be an easy fix, as I am very eager to try out this program of yours!
Regards,
Matt
DerStrom8, Wed Jul 23 2014, 12:55PM
Hi Sigurthr, thanks for posting.
I get an error when I try to run the exe on Windows 8:
I did some quick research and found this:
Hopefully this will be an easy fix, as I am very eager to try out this program of yours!
Regards,
Matt
Re: DRSSTC Peak Primary Current formula?
Sigurthr, Thu Jul 24 2014, 01:37AM
My Apologies, DerStrom8! I accidentally uploaded the debugging version instead of the release version.
Here's the correct link:
Sigurthr, Thu Jul 24 2014, 01:37AM
My Apologies, DerStrom8! I accidentally uploaded the debugging version instead of the release version.
Here's the correct link:
Re: DRSSTC Peak Primary Current formula?
DerStrom8, Thu Jul 24 2014, 01:48AM
That one works much better Sigurthr, thanks! Now I've got a stupid question: How would one determine the percentage to derate their capacitor by? Guess I'm not quite sure how to figure that out in this case.
Regards,
Matt
EDIT: I just assumed 60% for sake of argument and got the following for my Tesla coil:
Seems legit to me! Thanks a lot, this will help a lot!
Matt
DerStrom8, Thu Jul 24 2014, 01:48AM
That one works much better Sigurthr, thanks! Now I've got a stupid question: How would one determine the percentage to derate their capacitor by? Guess I'm not quite sure how to figure that out in this case.
Regards,
Matt
EDIT: I just assumed 60% for sake of argument and got the following for my Tesla coil:
Seems legit to me! Thanks a lot, this will help a lot!
Matt
Re: DRSSTC Peak Primary Current formula?
Sigurthr, Thu Jul 24 2014, 02:01AM
It's just an amount you want in the voltage overhead for the cap. I use 10% generally, but it really depends on cap type. If your MMC is rated for 30kV then you could probably go to 5% as that would leave 1500V overhead.
The reason I put the figure in there at all is that it is an easy way to adjust the outputs of the program without changing multiple inputs. For example: if you put in 0 derating and it says you can't run more than 500uS bursts without exceeding the cap rating, but the peak current is well over what your bridge can handle, for example maybe 1200A, you can re-run the program with various derating factors to find a new maximum usable burst length that doesn't exceed your bridge's current capability.
Sigurthr, Thu Jul 24 2014, 02:01AM
It's just an amount you want in the voltage overhead for the cap. I use 10% generally, but it really depends on cap type. If your MMC is rated for 30kV then you could probably go to 5% as that would leave 1500V overhead.
The reason I put the figure in there at all is that it is an easy way to adjust the outputs of the program without changing multiple inputs. For example: if you put in 0 derating and it says you can't run more than 500uS bursts without exceeding the cap rating, but the peak current is well over what your bridge can handle, for example maybe 1200A, you can re-run the program with various derating factors to find a new maximum usable burst length that doesn't exceed your bridge's current capability.
Print this page