| lgreen |
I've started this thread to post some measurements of my Gainclone. This is an LM3886 inverting GC with buffer and 3300 uF per rail, 500 VA transformer.
We see a lot of gainclones here but very little in the way of performance data. If you have some data, add your own.
Here is a shot of the onset of clipping with a 40 Hz input signal into a 5 ohm load, about 45 watts. 5 v/div on channel 1. |
|
|
| lgreen |
Here is the onset of clipping at 4 KHz into 5 ohms. Again, about 45 watts.
5 v/div on channel 1. |
|
|
| lgreen |
Here is the power supply ripple when the chip is muted. The power supply gives about 35 v DC.
Channel 1- 50 mv/div (since probe was changed to 1x) |
|
|
| lgreen |
Power supply ripple when the chip is operated and not clipping.
channel 1, .5v/div
It would be interesting to compare this to an implementation using 1000 uF per rail. I had one of these but added more caps due to the large amount of ripple. Now the ripple is not so bad, but still not great. |
|
|
| lgreen |
| Here is the positive power supply when the amp is clipping, and flattening the top peaks. This is a DC measurement at 5 v/div, and ground is the bottom line on the graticule. Thus you are seeing ripple at about 30-35 volts, much higher than the clipping point of about 12 volts in the first 2 posts. |
|
|
| moamps |
| quote: | Originally posted by lgreen
Here is a shot of the onset of clipping with a 40 Hz input signal into a 5 ohm load, about 45 watts. 5 v/div on channel 1. |
Hi,
Your scope reading doesn't look good. Did you check input waveform in GC? Clipping starts at 12V peak according to your picture, which is only ca 15Wrms/5ohm. How did you get 45W?
Regards
Milan |
|
|
| thomas997 |
What scope is that?
Hopefully I will get one soon and will add some pictures to the list. |
|
|
| rdf |
| Why the asymetrical clipping? Is it typical of these chips? It almost implies a serious DC offset on the output, but that would be hard to miss. |
|
|
| lgreen |
Yeah, the clipping is not symetrical, but DC offset is never more than 25 mv. I'll look at it again into different loads (maybe different results) but I would have no idea what's causing it, since it seems to be working fine. That's why I posted my data, wondering what everyone else's looks like on clipping?
Scope is a Tektronix 265M. |
|
|
| rdf |
| Are both PS rails the same voltage? |
|
|
| lgreen |
| 4 KHz square wave, output is the top trace, 1x probe for all these next pics. |
|
|
| lgreen |
| 4 KHz triangle wave, output is the top trace |
|
|
| lgreen |
15KHz square wave, output is the top trace.
The bottom trace is the input and it was not quite locked so you see a little movement show up in the picture. |
|
|
| lgreen |
| 15 KHz triangle wave, output is top trace |
|
|
| lgreen |
20 KHz square wave, output is the top trace.
Geeze, this GC doesn't roll off the high frequencies, does it? |
|
|
| lgreen |
40 KHz square wave, same deal.
I guess I never said it but in all these 2 trace pics, the bottom is the input and the top is the output. |
|
|
| lgreen |
this is the Low Noise power supply, shown here. This has the input buffer (differential receiver) on it.
In these shots I am looking for noise on the power supply. Scope is set at .5 msec/div (and its highest sensitivity which is 5 mv/div). This rail is +9.5 volts.
This is the positive rail with the amp operating but muted.
It also shows the entire scope for those of you interested in seeing my TEK 465M. |
|
|
| lgreen |
| Just to compare, this is the same thing with the amplifier turned off. Trying to see what the probes pick up when the amp is off as compared to when operating. |
|
|
| lgreen |
Here is the +9.5 volt rail of the low noise power supply input buffer with a 4KHz input signal applied. You see that I am getting a 4KHz input signal component on the rail. Dont know if this is a measurement artifact (picked up by the probes and wires) or if its actually there. Interesting though.
The scope is still at its highest sensitivity, 5 mv/div (top trace) where the bottom trace is just showing the input at a lesser setting.
This pic also shows the entire scope. |
|
|
| moamps |
| quote: | Originally posted by lgreen
......since it seems to be working fine.....
|
If you think so....
...Your scope looks nice, but some readings not...
(If you want to somebody help you, set your scope channels to DC, move second channel var input pot fully clockwise, set divs to 10V, traces position in midle, set sine signal 1kHz on input, increase level just to clipping and show as result; first channel to output, second on +rail...only one picture please...)
Regards
Milan |
|
|
| thomas997 |
| Why do the last few pictures look blurry on the scopes monitor? |
|
|
| roibm |
Showdown somehow?
I don't remember seeing any shots for the Brian's gainclone either. Would be interesting to get some, but I can expect that from the builders only ;) |
|
|
| lgreen |
| quote: | If you think so....
...Your scope looks nice, but some readings not...
(If you want to somebody help you, set your scope....
|
I realize this is clipping a little oddly, I suspect the reason is that my signal source had a positive DC offset for those shots that I took. I'll look at this again and post the results but I don't think I need any help just yet. Thanks for offering though.
| quote: | | Why do the last few pictures look blurry on the scopes monitor? |
You are referring to the power supply shots. These should be a straight line, but due to added noise they appear to be fat lines. They look fat because they are essentially showing a bunch of hash/noise that varies over time, so on a digital photo it looks like a really thick line.
| quote: | Showdown somehow?
I don't remember seeing any shots for the Brian's gainclone either. Would be interesting to get some, but I can expect that from the builders only |
Well, I don't think showdown, but lets see what everyone else is getting. I'm interested in frequency response, square wave response and power output, and whatever else you all think would be interesting. This way when someone builds one they can ensure that they are getting the results that they should be getting. |
|
|
| Sandy H. |
My old tek has a 2nd channel which does not work correctly most times. If there is a measurement which I could take using just channel 1, with a generator, I will post results for a point to point LM3875 chipamp I built last year. I bought the scope but never really learned to use it well.
If anyone wants to help and give me a set of guidelines to produce curves for comparison, my email is below. I should have an hour or two this weekend to play and would like to contribute.
Sandy. |
|
|
| PMA |
Why are you trying to blame everything on the scope? It is not that difficult to calibrate it.
I would rather admit that there might be something wrong with the amp - most probably both the chip and the design. |
|
|
| Sandy H. |
Just bumping the thread in case anyone want to help me know how to measure my GC with my tek 465 scope to get another data point. I've not used a scope before and the settings aren't clear to me.
Let me know!
Sandy. |
|
|
| tschrama |
| Since you invited people to submit their GC measurements....I once did some comparison THD tests. One of the amps I measured was a LM3875TF inverted gainclone. The THD residual of my oscilator is -103dB.. this picture show the measurements of the gainclone... |
|
|
|
