1n5399 vs 1n4007 for diode stack / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

1n5399 vs 1n4007 for diode stack

previous 1 2 3 4 next

Move Thread

LAN_403

Matt Edwards

Sat Aug 28 2010, 11:30PM

Registered Member #2838 Joined: Fri Apr 30 2010, 07:55PM
Location: tehachapi, CA
Posts: 333

Thanks for the advice Proud Mary. I was looking into buying some 30 or 60k diodes but it would be unfortunately way too expensive. I don't mind putting in the extra work if it will save me some money.

teravolt

Sun Aug 29 2010, 02:52AM

Registered Member #195 Joined: Fri Feb 17 2006, 08:27PM
Location: Berkeley, ca.
Posts: 1111

Patrick wrote ...

Proud Mary wrote ...

Wouldn't it be simpler, cheaper, and neater to just buy diodes of the appropriate rating?

yeah if he can find the ratings he needs with low enough cost.

And to Teravolt, in your above full wave CW pic, how much cumulative time of operation do you have on that setup? the longer the time you have logged the better the 1n4007's are sounding. How much power are you putting into those 4 flybacks (at low V) ? How much power do you estimate is leaving the flybacks at HV?

hello Patrick, the output from those flybacks are about 120ma peak as I mesured it with a current probe and O-scope.
hears what it looks like

it is quite a hot spark. You were making a reference to "hold back" do you think that a uf4007 would be more robust in a CW trippler around 66 khz than a standard 1n4007

Oidium 45 yesterday I blew the hv dubilier section that is in Pripyat's thread encased in wax trying to take a picture and am in the process of rebuilding it. so if you pot it a series resister with the output is a good idea.

hears Pripyat's thread for reference

Matt Edwards

Sun Aug 29 2010, 03:30AM

Registered Member #2838 Joined: Fri Apr 30 2010, 07:55PM
Location: tehachapi, CA
Posts: 333

teravolt wrote ...

Oidium 45 yesterday I blew the hv dubilier section that is in Pripyat's thread encased in wax trying to take a picture and am in the process of rebuilding it. so if you pot it a series resister with the output is a good idea.

hears Pripyat's thread for reference

I remember seeing this one. Was it the diodes that died?

Patrick

Sun Aug 29 2010, 05:34AM

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

teravolt wrote ...

it is quite a hot spark. You were making a reference to "hold back" do you think that a uf4007 would be more robust in a CW trippler around 66 khz than a standard 1n4007

Trr = reverse recovery time --or-- reverse recovery transit

well, I was a litte bit drunk when I wrote that. Maybe I should have described it better. "hold back" was all I could think of to describe the current during the Trr time. During the Trr time current is not blocked and continues to flow in the wrong direction before the charge carriers/holes have time to recover, thus with small caps you can lose current and voltage(since charge in a cap is proportional to voltage) so therefore a small Farad cap with a long Trr time diode looses way more power than a large Farad cap and a fast Trr time. So, i f you want the most from your caps you should use the fastest diodes possible/practical. With CW's (half or full) alot is up to the end users preference so there dont seem to be many definate rules, just relationships to be adjusted.

So, I was always taught that the diodes stay cooler with a short Trr time, and the ripple/power loss is much less. Since the capacitence of the caps is the most valuable part of deciding the throughput power, I always use fast diodes.

so, I would think that the UF4007's would be better, at least for engineering considerations. Cost is more though.

look here
from the above link:
50kHz = 0.00002 seconds (switching freq)
70nS =0.00000070 seconds (Trr)
therefore: 50khz/ 70nS = 28.5

so, according to VMI they advise a diode at least 28.5 times faster then the wave your retifying.

teravolt

Sun Aug 29 2010, 06:16PM

Registered Member #195 Joined: Fri Feb 17 2006, 08:27PM
Location: Berkeley, ca.
Posts: 1111

Hello Oidium45 the diodes definitely died and I am not sure how many out of 60. the caps I am not sure of either .Will report later. right now I am only half the way through the autopsy.

thanks for the info Patrick, If I understand the junction capacitance is responsible for the short time between blocking and conducting. When I was measuring the output frequency can very depending on a load. The output is sinusoidal at 66khz max, my coils are fairly small so a bigger coils may have a lower frequencies depending on flyback resonance. If the output was a square wave I would definitely use uf4007's but being that it is a sign wave the output's di/dt is not to fast for the diodes so it works.
That being said I am not sure whether uf4007 would give better or the same results or be more robust with out doing the experiment. I can't believe that it would be worse. It all depends on the other components and what you want to do with it and your buget.

Patrick

Sun Aug 29 2010, 08:49PM

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

teravolt wrote ...

Hello Oidium45 the diodes definitely died and I am not sure how many out of 60. the caps I am not sure of either .Will report later. right now I am only half the way through the autopsy.

when did it pass away? were you operating it to further this thread, when it died?

teravolt wrote ...

That being said I am not sure whether uf4007 would give better or the same results or be more robust with out doing the experiment. I can't believe that it would be worse. It all depends on the other components and what you want to do with it and your buget.

if your sure its sinusodal or near sinusodal then your right, mine is very close to square or trapazoidal, i think. ( I need my CVD to be sure. ) The reason why i like the fastest diodes possible can be imagined this way:

lets put a 20kVDC across 20 SPST switches all in series, now if we throw-closed some early and some late we would expect to see 3,4,5 kV across some (with junction, lead L&C)and a full 20kV across the unlucky last to close bugger. if these swithces were instead 1kV diodes, you can see the problem.

wether its 1.0uS or 75nS as long as there all the same they will have an average distrubution of early and late Trr, but with the faster diodes the rise time to exceed breakdown doesnt last as long, so faster diodes may be somewhat more reliable.

(it should be noted this expalanation is over simplified, there are some factors that allow the diodes to "cooperate" in thier switching times.)

Also, its more then just junction capacity that determines Trr. But thats part of it.

teravolt

Mon Aug 30 2010, 07:36PM

Registered Member #195 Joined: Fri Feb 17 2006, 08:27PM
Location: Berkeley, ca.
Posts: 1111

when did it pass away? were you operating it to further this thread, when it died?

yes, it died during this thread but I was getting a stream 2" of blue arcs befor they crapped. I be will adding some resistance to cut the current. They were fine at idle but I think the discharges is what killed them. I do not think that 1n4007 Trr is rated for that,

Matt Edwards

Mon Aug 30 2010, 08:00PM

Registered Member #2838 Joined: Fri Apr 30 2010, 07:55PM
Location: tehachapi, CA
Posts: 333

I have decided to chip out for the SF1600's. Unfortunately there is a two month wait for just about every supplier that I have checked. Guess this project will have to wait a bit.

Patrick

Mon Aug 30 2010, 08:43PM

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

Oidium45 wrote ...

I have decided to chip out for the SF1600's.

use an output resistor to limit the current to save your precious SF1600's as Teravolt suggests.

From this thread with what others and Teravolt have said, i may try some cheaper less capable diodes then my usual HER's or SF's, for my next CW. Once I get my CVD up and running ill be able to figure stuff out, like quatifying and qualifying certain factors versus load conditions.

Matt Edwards

Tue Aug 31 2010, 01:02AM

Registered Member #2838 Joined: Fri Apr 30 2010, 07:55PM
Location: tehachapi, CA
Posts: 333

Patrick wrote ...

From this thread with what others and Teravolt have said, i may try some cheaper less capable diodes then my usual HER's or SF's, for my next CW. Once I get my CVD up and running ill be able to figure stuff out, like quatifying and qualifying certain factors versus load conditions.

Let me know how it works out for you!

Thanks again everyone for the input!!!

previous 1 2 3 4 next

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