• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Scope tricks for detecting rectifier/choke switching

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Was wondering if anyone has some ideas on how to best detect both rectifier switching artifacts, as well as choke ringing when commissioning your power supplies.

Was going through Morgan Jones 3rd ed, and he has some pictures of examples of choke ringing, which were helpful to an extent. It's easy to look at the transformer secodary and see if there is any ringing, which I've been successful snubbing with R-C. Detecting subtle aberrations in the choke voltage or rectifier switch-off is a lot more difficult.

Problem is I have a hard time (20Mhz analog scope) getting a good zoomed in view of the waveform to see rectifier switching, and choke ringing. I've played with the trigger level and timebase to get it as close as possible, but just can't seem to get as close as I could with my (now sold) TDS210 digital scope. When I get close enough, I lose the trigger.

I am up a creek without a paddle?
 
zigzagflux said:
Was wondering if anyone has some ideas on how to best detect both rectifier switching artifacts, as well as choke ringing when commissioning your power supplies.

Was going through Morgan Jones 3rd ed, and he has some pictures of examples of choke ringing, which were helpful to an extent. It's easy to look at the transformer secodary and see if there is any ringing, which I've been successful snubbing with R-C. Detecting subtle aberrations in the choke voltage or rectifier switch-off is a lot more difficult.

Problem is I have a hard time (20Mhz analog scope) getting a good zoomed in view of the waveform to see rectifier switching, and choke ringing. I've played with the trigger level and timebase to get it as close as possible, but just can't seem to get as close as I could with my (now sold) TDS210 digital scope. When I get close enough, I lose the trigger.

I am up a creek without a paddle?


My scope is very similar to yours and the switching pulses occurr every 20ms (half wave).
I killed them with a 100nf across the transformer winding.
The switching pulses could be heard through my tweeters.
 
I guess I'm not stating my goal clearly enough. Since I don't want to break any copyright laws, I won't scan and attach a copy of page 316 of Morgan Jones' 3rd edition. The bottom graph (Figure 5.16) is really what I'm trying to capture.

I want to play around with C's or RC's to get that little transient eliminated. I am reasonably skilled with my o-scope, and I understand the above suggestions, but I'm just not able to get the same type of resolution.

Maybe it's just the scope. I am considering splurging for a color 4-channel Tek, but need to wait a few months yet.

Note also his success story between Figures 5.17 and 5.20. I don't see a whole lot of difference between the voltages: it's the current that really shows the improvement. I don't see a good way to capture the current without an active probe. Would subtracting channels 1-2 across a resistor really work well enough? Got to break solder connections to get that.
 
How about something like this?

http://www.aemc.com/products/pdf/1200.67.pdf

Only reason I picked it was the low mA measurement ability; the 2 kHz bandwidth isn't very desirable, especially since I'm looking for switching transients.

Other option would be the Fluke 80i-110s, which is good for 100 kHz, but only reads (on a practical level) to 10-50 mA. Might be a little noisy this low in its measurement range.

http://assets.fluke.com/datasheets/80i-110s-Spexs.pdf

What would be a good bandwidth to shoot for to detect and correct rectifier switching transients in the current waveforms? I can get some very high bandwidths with Tektronix probes, but those are in excess of $2000.
 
zigzagflux said:
How about something like this?

http://www.aemc.com/products/pdf/1200.67.pdf

Only reason I picked it was the low mA measurement ability; the 2 kHz bandwidth isn't very desirable, especially since I'm looking for switching transients.

Other option would be the Fluke 80i-110s, which is good for 100 kHz, but only reads (on a practical level) to 10-50 mA. Might be a little noisy this low in its measurement range.

http://assets.fluke.com/datasheets/80i-110s-Spexs.pdf

What would be a good bandwidth to shoot for to detect and correct rectifier switching transients in the current waveforms? I can get some very high bandwidths with Tektronix probes, but those are in excess of $2000.

I have a 20MHz scope and that captures rectifier switching very well.

I built my own digital scope many years ago using a PIC with a memory chip. It could either capture 8 channels simultaneously or a couple of analog channels simultaneously.
Once it was triggered and captured the memory full of data it was uplaoded via the serial port to a PC program that displayed teh results.
It reall ywasnt tha thard.
 
Some progress

Okay, bought the scope and current probe, started to poke around. This is fun (and a little dangerous). Pictures attached.

I have a capture of the transient in the current. Just as MJ and SY noted, the transient is absent from the voltage, yet most noticeable on the current. I would like to eliminate this transient.

There doesn't seem to be much ringing, though that may have to do with the bandwidth of the current probe (-3dB at 100 kHz).

Slugging what remains has proven a little tricky, but I'll keep you posted as to the results.

http://www.just4sheep.com/RevRec.pdf
 
There's some nice progress. Tried out caps from 0.1 to 0.47 uF directly across the PT secondary. Values from 0.22 to 0.47 resolved the transient, but the larger the cap got, the more grunting I could hear out of the transformer.

Reduced the cap to 0.22 uF, which seemed to be the minimum needed to clear up the transient. Then, I proceeded to place resistors in series to ease up on the PT, while still clearing up the transient. 100 ohms worked, but you could still see a slight hump. 50 ohms was about the right value. Switching the snubber in and out, you can't hear any audible difference in transformer noise. Time to get some carbon comps on order.

Pictures attached showing the improvement. Just to be redundant, I would point out that at no point (snubber, no snubber) was a change in the voltage detected. It was only by watching the current that I could see any progress. Hope this results in a nice clean supply at the end of the day.

http://www.just4sheep.com/site/images/fixed.pdf
 
It's my lucky day - your power supply matches mine right down to the rectifier parts!

Thank You for posting such clear & precise measurement, I'll use these snubber values & compare (I'm using something like 500pF now).

What cap construction did you try? I've got some 1000V X2 polypropylene box-types at this value.
 
Rod:

Hope it works for you. Only caveat I would have for you is that the transient is highly dependent on the inductance of the transformer secondary; I am using an Electraprint custom. If yours is different, Your Mileage Will Vary.

I am a huge fan of the Cornell Dubilier 940C series; they are double metallized polypropylene, and have insane ESR and current ratings. I use them for both power supply and snubber caps. My 0.47uF cap immediately after the rectifiers uses this cap as well.

http://www.cde.com/catalogs/940C.pdf

I was hoping to use an orange drop, but their AC ratings are quite low in comparison. Might want to double check that X2 capacitor for it's AC rating; I went with 1600VDC/630VAC caps throughout.

Good Luck !!
 
oshifis:

Are we thinking of the same thing? I was trying to snub the high frequency transient that occurs when the rectifiers switch off. The transformer doesn't want to stop its current flow, so spits out a back EMF. The RC network damps this energy.

With regards to a capacitor in parallel with the choke, I actually have one already; it's the series combination of 0.47 and 120 uF, which is 0.47 uF. If your suggestion is to pick 120 = 1 / (LC)^0.5, I have a frequency of 460 Hz.

Could you clarify?
 
Not exactly. I mean the following filter topology: 0.47u to GND, 10H and 0.18u parallel forming a tuned circuit, then 120u to GND. The only difference to your circuit is a 0.18u tuning capacitor parallel with L. This in effect form a low-pass elliptical filter (Cauer, I think). With proper tuning of L and C (10H and 0.18u) it has infinite attenuation at 120 Hz, and some quite high stopband attenuation above. It has nearly zero additional cost, just an extra capacitor.
 
Well, you can simulate it yourself. I found the free evaluation version of Tonne Software's Elsie that can simulate such filters. Never tried it in practice, though. Initially I also believed that the capacitor before the choke and the one after it are connected in series and resonate with the choke. This is not the case, they act independently and a separate capacitor is needed across the choke to resonate at 120 Hz.
 
oshifis said:
Initially I also believed that the capacitor before the choke and the one after it are connected in series and resonate with the choke. This is not the case, they act independently and a separate capacitor is needed across the choke to resonate at 120 Hz.

I guess my point is that this is exactly the case, they are in series. I fail to see how a separate cap in parallel is any different. As well, the transient I have snubbed out would happily pass through this parallel capacitor, as opposed to being blocked by it. The goal is to damp this transient to where its energy is absorbed. Blocking the energy is not the idea. I really don't think we're talking about the same goals.

I am aware of people installing RC filters across the choke, to reduce choke ringing. MJ addresses this in his book, but claims an improved method with C's just before and after the choke (again, a very similar solution, but without the R). I think this still does not address transformer back EMF or stored reverse recovery charge.
 
The resonant capacitor / choke was commonly used in the early AC radios of the 1930s... it gave adequate suppression of 120 Hz with the small values of filter caps that were available then. Once electrolytics were developed, it was seldom seen. But the Radio Amateur's Handbook mentioned it, as I recall... you have only smaller value oil capacitors with the high voltages used in transmitters.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.