So just like @east electronics I've been messing around with some P3A's and I had one lying around for some 10 years or so, then I finished it recently but there is one odd problem. When I test it with a square wave of anything between 100hz and above I get a output that is a very spiky overshoot which is as vertical as you can get and then a slow ringing ramp up, on every cycle. I attached a hand drawn approximation of what my analog scope shows.
I have everything as in the schematic from Rod's page, 220pF WIMA PP then 100pF WIMA PP for dampening at the drivers. In the output is the original 10 ohm series with 100nF high frequency filter .
I also have WIMA 100nF as bypass at each rail close to output transistors.
I had a 100uF parallel with the 100nF , took those out no change.
I also tested both with 8ohm resistor dummy load and just scope probes alone and again no difference.
At first I thought it might be too long PSU rail wires etc, but taking adding the 100uF electrolytics on board changed nothing.
One last idea I have currently is to disconnect feedback and see whether the problem is at the output or it comes from the input because with feedback attached the input also shows a minimized output waveform as it comes through the feedback loop.
Can you guys suggest any ideas?
I have everything as in the schematic from Rod's page, 220pF WIMA PP then 100pF WIMA PP for dampening at the drivers. In the output is the original 10 ohm series with 100nF high frequency filter .
I also have WIMA 100nF as bypass at each rail close to output transistors.
I had a 100uF parallel with the 100nF , took those out no change.
I also tested both with 8ohm resistor dummy load and just scope probes alone and again no difference.
At first I thought it might be too long PSU rail wires etc, but taking adding the 100uF electrolytics on board changed nothing.
One last idea I have currently is to disconnect feedback and see whether the problem is at the output or it comes from the input because with feedback attached the input also shows a minimized output waveform as it comes through the feedback loop.
Can you guys suggest any ideas?
Attachments
Member
Joined 2009
Paid Member
In addition to oscillations, check that it isn’t shoot-through in the output stage. I had this happen on mine and it’s cure is to increase standing current through the drivers.
My exploits are here: https://www.diyaudio.com/community/threads/tgm8-an-amplifier-based-on-rod-elliot-p3a.245619/
Look at posts starting #238
My exploits are here: https://www.diyaudio.com/community/threads/tgm8-an-amplifier-based-on-rod-elliot-p3a.245619/
Look at posts starting #238
I might need it yes.
Well I've built quite few over the years but only have tested few of them. I actually got back to them because as I tested one older build I found out that although it's stable etc but it doesn't produce a nice square wave at the output , the square wave was sort of triangular but it had no overshoot or anything like that. I disassembled it made a better PSU, new caps etc, changed the ceramic stability 100pF caps to WIMa's etc but have not tested it yet, I started with this single channel single board older piece, added same new caps and all and then found out that instead of being sort of triangular when input a square wave it now does this overshoot spike and ringing.
I don't think it's shoot through , there seems to be some dead time in the scope, my bias current is about 20 to 30mA depending on whether no input signal or whether there is input static.
Yes , the input signal is fine. It's a PC sound card, and software tone generator. I checked the sound card output with the scope and it shows a nice and clean square wave.
Now The output transistors I use for this one are the former Fairchild now OnSemi FJL4215 and FJL4315
What is weird is the vertical spike which then dips just as fast as it originates and then the waveform starts as a somewhat linear 45 degree ramp which rings along the way. The slow ramp up of the main cycle part would indicate a inductance but why then the sharp vertical spike?
What is the cycle time (frequency) of the signal?
Looks like it has damped highs. Check for the caps values if there is bigger value mistaken for smaller.
Check also if there is wrong 100nF connection, connected to the wrong point.
And precise schematic (picture) can be great help for troubleshooting.
Looks like it has damped highs. Check for the caps values if there is bigger value mistaken for smaller.
Check also if there is wrong 100nF connection, connected to the wrong point.
And precise schematic (picture) can be great help for troubleshooting.
PC soundcard connected to an amplifier=gnd loops.
Take care and measure the input signal with amplifier connected or use a stand alone AF generator if available.
Take care and measure the input signal with amplifier connected or use a stand alone AF generator if available.
The initial edge would propagate through the parasitic capacitance across/in parallel to the inductor. After the edge, you would get a ramp as the current through the inductance starts to increase. The things you see are consistent with an inductance, but it doesn't have to be a physical inductor.
Anything that has an impedance that increases with frequency acts as an inductance, like the output impedance of an amp.
Jan
Ok folks, so here is the problem it seems. For whatever reason my soundcard months before gave me a clean square wave output, I still recall it, but now it gives me a total garbage, so currently I am on hold with regards to the amplifier as I need to get myself a decent signal generator.
My sound card is old though (Creative audigy 2 zs or somethin like that).
I tested this with two scopes and both showed the same stuff so I can almost definitely say the card is bad.
It also does this with two different RCA cables.
I suspect the amplifier should turn out fine after all because it was able to replicate that very sharp vertical spike that it had on it's input from the sound card.
Anyway, I will get myself a signal generator which works and then test the amp once more and come back with the results.
By the way I have attached a picture of what my sound card outputs as square wave now.
My sound card is old though (Creative audigy 2 zs or somethin like that).
I tested this with two scopes and both showed the same stuff so I can almost definitely say the card is bad.
It also does this with two different RCA cables.
I suspect the amplifier should turn out fine after all because it was able to replicate that very sharp vertical spike that it had on it's input from the sound card.
Anyway, I will get myself a signal generator which works and then test the amp once more and come back with the results.
By the way I have attached a picture of what my sound card outputs as square wave now.
Doesn't your scope have a 1kHz square wave test signal source (usually it's for testing and where possible trimming the probes to produce a good looking square wave), you could maybe use that signal source to test your amp but may need to be attenuated, and perhaps time to update the scope too, what model is it?
It looks like his scope is working pretty good! Changing to a newer digital scope is usually a downgrade for audio use.
No it doesn't have such feature, this is an older Soviet scope named C1-94 or on english it would be S1-94 , but within it's limits it works fine, it is definitely good enough to test a P3A amp, one downside is that it doesn't have good synchronization therefore it is sometimes hard to catch certain frequencies on the screen for them to be clearly visible.
Why would a good digital scope be a downgrade? You think the sampling wouldn't show some high frequency noise on top of the sine or something else?
A CRT gives a lot better view than a LCD. Those fuzzy lines on your square waves would look a lot different.
True but I do like the better sync of the newer scopes and memory and other features, on an older analog one I basically have to catch the artifact I'm lookin for otherwise it's gone
So I got myself a signal generator, nothing fancy, the Uni-T UTG962E version and did a quick test and saw that the output is fine.
I did not manage to get to high frequency square wave because my old CRT scope has some synchronization issues but other than that at some 300-500hz it seemed fine, sharp edges, no overshoot , I guess I can call it a day.
The problem was definitely the sound card producing some bad output.
I did not manage to get to high frequency square wave because my old CRT scope has some synchronization issues but other than that at some 300-500hz it seemed fine, sharp edges, no overshoot , I guess I can call it a day.
The problem was definitely the sound card producing some bad output.