Suggestions for a low-interference smpsu please. / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Suggestions for a low-interference smpsu please.

1 2 3 next

Move Thread

LAN_403

Sulaiman

Wed Oct 05 2011, 11:21AM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

I'm (very) slowly designing/building a hf transceiver
a part of it needs to be an inverter

Input : 12 V lead-acid battery, VRLA or automotive.
Output : 28 Vdc at up to 5 Adc for the hf Power Amplifier.
(ssb, 'speech' power demand)

I want to minimise electromagnetic and electrical interference to the rest of the transceiver.
I considered 'royer' types but the component values are impractical,
so I'm going for a hard-switched push-pull forward converter

I found a nice big Mullard pot core
(FX2243..don't look for data on this modern internet)
and some Litz wire for winding, I've got a couple of 80V 90A TO220 mosfets

and a dual schottky rectifier and TL494s etc.
(may change from TL494)
The output inductor will be a gapped pot- or EI-core
and I've got some electrolytic capacitors on order.

Anyone have any experience with low-noise smps?
I need hints, tips and tricks!
(I've read quite a bit)
e.g. I thought maybe a 32,768 Hz fixed frequency might help identify spurious frequencies, or maybe a frequency offset pot to move 'spurs'....
It's only a small part of the transceiver but possibly a source of difficult problems.
The inverter should fit on a 100 x 160 mm card in a 19" rack.
Hand-wired on a matrix board, so not too complex.

Any hints, tips or tricks?

Patrick

Wed Oct 05 2011, 02:21PM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

I think what you want is a fixed on time variable off time, so PFM. that will make the noise easier to isolate.

WaveRider

Wed Oct 05 2011, 07:56PM

Registered Member #29 Joined: Fri Feb 03 2006, 09:00AM
Location: Hasselt, Belgium
Posts: 500

Hi Sulaiman.. Perhaps the best way to reduce noise on the output of an SMPS is to set the SMPS output voltage just above your required voltage and follow the SMPS with a low-dropout series pass regulator. This can provide in excess of 60dB of noise isolation as the LDO strips the ripple off the SMPS output.. The only drawback is the efficiency loss through the series regulator. However, the alternative is using thumping big inductors and capacitors which will decrease efficiency anyway, as well as destroying the output voltage regulation because of inductor series resistance.

If you are building the transceiver, use a further layer of regulation on all frequency sensitive subsystems (PLLs, VCOs, etc.). If you can tolerate the efficiency loss, an extra layer of regulation on your transmitter final amplifier will keep power supply spurious out of your transmit signal.

Have fun!

Sulaiman

Wed Oct 05 2011, 08:25PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

Patrick, I get the 'fixed on time' PFM scheme, thanks, I'll consider it ..... which ic does pfm?

WaveRider, a 5A ldo linear regulator, good idea but probably a bit too much stuff for this project but I'll keep it in mind if required. thanks.

Patrick

Wed Oct 05 2011, 10:06PM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

Sulaiman wrote ...

Patrick, I get the 'fixed on time' PFM scheme, thanks, I'll consider it ..... which ic does pfm?

Shit. I was afraid youd ask me that!

Ill look then post here again, GTG to school noww.

2Spoons

Thu Oct 06 2011, 01:22AM

Registered Member #2939 Joined: Fri Jun 25 2010, 04:25AM
Location:
Posts: 615

Look at some of the Cuk topologies. There are some clever versions using coupled inductors that can give you zero (yes really!) input or output ripple. Also ZETA for very low output ripple, or SEPIC for very low input ripple.
If its harmonics that concern you you might try a quasi-resonant approach.
IF you are interested in only a small portion of the HF band then careful selection of operating frequency can put the harmonics outside the area of interest - and also outside the IF frequency. I used this trick to good effect when picking a microprocessor crystal for a GPS tracker I designed.

WaveRider

Thu Oct 06 2011, 07:52AM

Registered Member #29 Joined: Fri Feb 03 2006, 09:00AM
Location: Hasselt, Belgium
Posts: 500

A search of the Farnell site yields this e-z to use 5A LDO with 1V dropout... (atho' 28V is pushing the limit on this device)

]lt108x.pdf[/file]

Since you are designing the transceiver from the ground up, a dedicated power supply with multiple regulators that provide 3.3V, 5V, 28V, and any required negative voltages is recommended. This can be powered by your 12V battery.

radiotech

Thu Oct 06 2011, 07:27PM

Registered Member #2463 Joined: Wed Nov 11 2009, 03:49AM
Location:
Posts: 1546

Two comments. The transceiver has lower power requirements on receive.
On transmit your RFI concerns are less.

Make two power supplies. low, easily shielded.
High. Less shielding.

When transceiver sends, switch both on.

History, old military units kicked in a dynamotor on transmit, otherwise operated on
a vibrator pack.

2Spoons

Fri Oct 07 2011, 12:29AM

Registered Member #2939 Joined: Fri Jun 25 2010, 04:25AM
Location:
Posts: 615

WaveRider wrote ...

. Perhaps the best way to reduce noise on the output of an SMPS is to set the SMPS output voltage just above your required voltage and follow the SMPS with a low-dropout series pass regulator. This can provide in excess of 60dB of noise isolation as the LDO strips the ripple off the SMPS output..

Linear regs are generally crap at rejecting signals over 10kHz, unless you can find something a bit special. The GBW product just isn't enough to provide any decent attenuation at MHz frequencies. Careful design of the SMPS is required first, to minimise generation of interfering signals - not just harmonics of the switching f, but also parasitic ringing, which can be pretty bad in some topologies.
That datasheet you posted shows rejection at 100kHz is down to about 25dB, and falling at ~20dB/decade.

Patrick

Fri Oct 07 2011, 01:01AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

2Spoons wrote ...

I was about to say the same thing. I think a good design will be most of the battle.

1 2 3 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob