Wierd clipping & barking at slightly low line (110VAC)

[nosaj]

Today, finally back on this odd clipping behavior, I fired up my Roland V-Sessions TD-10W controller, dialed up a decent low-pitched kick drum, and used that with an 8-bar pattern on a drum kit, while only feeding the kick and a bit of snare as the signal source, so this is closer to how this amplifier is used (apart from not having the drum throne with the seat transducer here).

Looked at it both at 120VAC and 110VAC. Same problem. At 110VAC, it comes unglued. The waveform distortion's scalloping of the output waveform appears to be in sync with the AC mains charging current.

[ATTACH=CONFIG]54172[/ATTACH] [ATTACH=CONFIG]54173[/ATTACH] [ATTACH=CONFIG]54174[/ATTACH] [ATTACH=CONFIG]54175[/ATTACH]

I also looked at it with resistive load, and get the same behavior. I've also attached the impedance plot of my test speaker.

[ATTACH=CONFIG]54176[/ATTACH] [ATTACH=CONFIG]54178[/ATTACH]

It's odd that with 60Hz and higher signal frequency, this doesn't happen, even at lower line.

[ATTACH=CONFIG]54177[/ATTACH]

I checked the ESR of the two 10,000uF/63V supply caps on this amp, and they measure 125mOhms @ 40Hz thru 1kHz, using a 20V P-P square wave, 50 ohm source, measuring the resultant voltage step in the waveform at the test frequency.

So far, increasing the current thru the voltage gain stage to 9mA (from 2.4mA), and 15mA thru the driver stage (up from 4.2mA) hasn't shown any change in this behavior. Mostly there's a trend of higher charging current once below 60Hz signal, and at 110VAC, with the waveform aberrations being in relative sync with the charging frequency. Oddness.

Any chance this is just the way it's designed for 120v?
I know some PA's and other things act real odd when starved for voltage.
Notice you have listed the brand or model or I missed it.


nosaj

[The Dude]

......I checked the ESR of the two 10,000uF/63V supply caps on this amp, and they measure 125mOhms @ 40Hz thru 1kHz, using a 20V P-P square wave, 50 ohm source, measuring the resultant voltage step in the waveform at the test frequency......

I'd pull and check the value of those caps. Let's say they are now 2,000uF instead of 10,000. You might still get decent ESR readings, but capacitance is low. 125mOhms seems high to me anyway, but I'm not familiar with your methods.

[nevetslab]

Any chance this is just the way it's designed for 120v?
I know some PA's and other things act real odd when starved for voltage.
Notice you have listed the brand or model or I missed it.


nosaj

Porter & Davies BC-2

[nevetslab]

I'd pull and check the value of those caps. Let's say they are now 2,000uF instead of 10,000. You might still get decent ESR readings, but capacitance is low. 125mOhms seems high to me anyway, but I'm not familiar with your methods.

You're right. I even had some used 10,000uF 80V 'pulls' from a Mackie SWA 1801 Powered Sub measure lower in the ESR, same method. As these are well glued into place, I was sort of avoiding the obvious, now that there's a trend in the AC mains current waveform suggesting that too. I suppose I could tack in a pair first. I'm still waiting for the designer to get back to me on this.

[nevetslab]

You're right. I even had some used 10,000uF 80V 'pulls' from a Mackie SWA 1801 Powered Sub measure lower in the ESR, same method. As these are well glued into place, I was sort of avoiding the obvious, now that there's a trend in the AC mains current waveform suggesting that too. I suppose I could tack in a pair first. I'm still waiting for the designer to get back to me on this.

Today I've measured the existing filter caps, 8700uF, with ESR a touch higher...170mOhm. I've ordered new United Chemicon parts that call their SMH series 10,000uF/63V part way lower ESR. I'll have to wait until middle of next week to receive those. Meanwhile, I clipped on some 15,000uF caps, raising the measured total to 23,000uF per rail. Didn't make any difference though.

I heard back from the designer, where he and one of his staff have done some digging into the issue, and haven't been able to recreate the problem. Now, they never test below 100Hz....he didn't specify what they use as a test signal, where it's a mic'd kick, or something like the Roland TD-10W controller I have here, or just steady state sine. As they've shipped well over a thousand of these with the stock circuit values, regardless of what we speculate on where it looks like they're starving the outputs and drivers, none of those changes have yielded any cure.

So, I've restored the original circuit values, other than having faster drivers installed and the heat-sink mounted bias xstr. Using 100Hz sine-random tone, still goes whacko at 8% low line. Changed back to the Roland TD-10W kick pattern, which I had tuned it's fundamental down to around 43Hz, that too goes whack at low line, but is fine at nominal line voltage.

I thought I had found the cause earlier today, after their designer asked about any signs of overheating on the xfmr secondary connections to the board. The solder joints on the header (Lead-Free Solder) looked cold and poorly, so I re-soldered those, along with the bridge connections and the filter cap connections. Same thing.

I just received one of their Drum Thrones to use as the load, it being what the amp is designed for, but will have to wait until next week to get my hands on one of our BC2 amps for comparison. Ain't physics fun???!!!!

[nevetslab]

Today I've had a Porter & Davies Drum Stool connected to this BC2 amp, still driving it with the Roland TD-10W Kick Drum, which I had selected from a number of drum kit sounds, editing the kick for a low-pitched fundamental, and only feeding that sound with a little bit of snare to the amp. After finding the same waveform aberrations at 110VAC, and NOT at 120VAC, I began looking at the DC voltage level when the clipping occurs. Seems like 35VDC was the threshold, though in all cases, there was AC mains current charging the supplies.

Wondering if this was just a voltage supply issue, I removed the load, so there wouldn't be any significant mains current flowing. I had the mains turned down to 80VAC, with the supplies being less than +/- 30VDC (didn't measure them, other than seeing the level on the scope), and no distortion aberrations in the waveform. But, while I was turning the mains voltage back up, still triggering on the output signal, I got that distortion again, still in sync with charging current from the mains. So, this isn't a Supply Voltage issue. I had wondered what would happen if I substituted the existing AC Secondary winding with that from a similar-rated amp's power xfmr. I've yet to do that, but bit by bit, eliminating causes.

There is a protection circuit between the power amp output and the output connector that interfaces with the drum seat. It's an array of incandescent lamps, all in parallel and the two banks wired in series using Thermistors to couple them. The time constant for them to start lighting up is around 7 seconds. With a 4 ohm dummy load, I was able to get the lamps lit up, after which there was no more waveform distortion (with the output level now being lower). Today, leaving the output high, low line, I never did get the lamp protection circuit light up.

[nickb]

I'd focus on those two 100uF caps that filter the input diff amps. If there were a power supply noise rejection issue then that input stage is the sensitive area. Just parallel them with 100uF and see what happens.

[nevetslab]

I'd focus on those two 100uF caps that filter the input diff amps. If there were a power supply noise rejection issue then that input stage is the sensitive area. Just parallel them with 100uF and see what happens.

As those were quite accessible, I replaced the originals with some low ESR types, but, still the same issue. Good thought.

Earlier, I moved the output ground connection from what seemed a higher impedance path to scraping some solder mask away between the large copper area between the main filter caps (ground), to see if it had anything to do with where load return current was flowing. No change, though.

[g1]

Any chance this is just the way it's designed for 120v?
I know some PA's and other things act real odd when starved for voltage.

will have to wait until next week to get my hands on one of our BC2 amps for comparison. Ain't physics fun???!!!!

Have you had a chance to test another unit of same model yet? I'm inclined to agree with nosaj that it may well be a design issue. Is this the fault that was the original complaint?

[nickb]

As those were quite accessible, I replaced the originals with some low ESR types, but, still the same issue. Good thought.

Earlier, I moved the output ground connection from what seemed a higher impedance path to scraping some solder mask away between the large copper area between the main filter caps (ground), to see if it had anything to do with where load return current was flowing. No change, though.

I had meant to type 1000uF, not 100uf. Try the bigger ones.

[nevetslab]

I had meant to type 1000uF, not 100uf. Try the bigger ones.

My new 10,000uF/63V United Chemicon buss caps arrived, so I removed the existing caps from the PCB, found the snap-in leads on these new parts substantially shorter, so not much left to fold over on the foil side, relative to what was removed. These measured 8960uF, ESR around 63mOhms, DF around 0.44 (120Hz measurement). What I had removed measured 8600uF & 8700uF, ESR around 95mOhms, and DF around 0.74. So, the 100Hz Square Wave capacitance measurement (of 8700uF) was pretty close, that being an in-circuit measurement.

However, this didn't make any difference with the scalloping of the output waveform in sync with the AC mains charging current at 8.3% low line (110VAC). I clipped on a pair of 1000uF/50V caps across the 100uF/50V buss caps for the front end circuit, and that made no difference either.

I also clipped in a dual gang 100K Helipot across the 47k resistors that feed the emitter resistors of the two NPN/PNP Diff Pair xstrs on the front end, dropped that value down to as low as 27k in parallel with the fixed 47k values....which increases the current of the front end, the gain stage and driver stage currents, though that too made no difference.

I tacked in a pair of signal diodes, reversed-pair parallel between the input bases and the feedback side bases of the front end. That too made no difference.

Increase the AC mains back up to 120VAC, and all is just fine.

I've been using the Porter & Davies Drum Seat (not sure which model I have without looking), and have been using the Roland TD-10W kick drum, having edited the kick for deep shell, weight damping, de-tuned the head down to around 50hz, and have it playing an 8-bar pattern so it's all repeatable (just Kick with a touch of snare), and as close as I can get to normal use of this amp/seat while being convenient here in the shop.

The waveforms still look like what I've posted using that Roland kick. To quote W.C. Fields....."It Baffles Science"!

[nevetslab]

Have you had a chance to test another unit of same model yet? I'm inclined to agree with nosaj that it may well be a design issue. Is this the fault that was the original complaint?

I'm still waiting for the return of one of our inventory P & D BC2's from a client...supposed to be sometime this week. I too am beginning to think I'll find this same behavior with another unit. If I don't, my next move will be to bypass their output protection lamp network, which I haven't done. I've only seen that come on after around 10 seconds, driving a 4 ohm dummy load. When it does begin conducting, the lamps rob a good deal of the output, and the odd waveform distortion aberrations disappear. Driving their seat with it's transducer, I've NOT seen the protection network light up.

It's NOT the fault that my client had sent it over for. They were claiming both it NOT producing output and was humming. Neither of which I found to be the case.

Still, just really peculiar that just dropping the mains 8% will get you into this behavior. Now, in defense of P & D, my butt doesn't feel anything wrong when I see the waveform getting chipped away, sitting on their seat being driven hard. Put that instead thru a woofer, the growling you hear from the amp/power supply sounds like it's going to melt down any moment! That's what got me chasing this in the first place.

[nevetslab]

With egg on my face, NOT having looked at the input signal all this time, having seen the bipolar LM7812T/7912T IC regulators and their output voltages still at +/-12VDC at 110VAC, I didn't think to look at the input. Turns out there isn't sufficient AC secondary voltage to run these at low line voltage. 23VAC C/T @ 120VAC, which is just 11.5VAC per half. Now, odd thing is, on the unregulated supplies, they're NOT equal.....14.6VDC and -15.3VDC. I don't have the circuitry for their preamp/signal processing, but, I'd expect to see at least +/-18V to +/-20VDC on the inputs to these regulators. That waveform distortion begins occurring at 115VAC. I've messed with the input level, EQ control and the output level controls, but can't get enough drive to get full output at low line, without having the preamp crap out. I've replaced the input filter caps (470uF/25V) with 470/35's, not making any difference.

I don't have any low drop-out regulators on hand, though that will only buy me a little bit of margin. I found LM2990T Pos 12V regulators and LM2940T Neg 12V regulators, where the drop-out voltage is 0.6V instead of 1.1V. Not much difference.

I found the difference on the unbalanced positive supply....they're running the output relay off of +12VDC, so it's dragging that unregulated supply voltage down further. That explains why I'm mostly seeing positive side clipping.

Back to chatting with the designer.

And, restored all of the minor changes I added.....restoring their slower drivers xstrs, restored the free-air bias xstr, while leaving the new filter caps in on the main board, as well as the preamp board. Ordered LDO regulators and see if they make any difference. Only real simple fix is to have their power xfmr revised for, say, 30VAC for the regulators.

Having just heard back from the designer, they can drop the mains voltage down to 200VAC without any of this action occurring, so it may be an earlier revision of the Toroidal Xfmr.....sending that data back to him in the morning.

[nevetslab]

Impedance Plot of Porter & Davies BC Gigster Drum Stool

This morning, having remembered how to set up my cobbled Impedance Plotting System (using 10mA Constant Current Sine Wave source), and getting resistance calibration plots from my ESI Decade Box of 2 ohm, 4 ohm, 8 ohm, 16 ohm, 32 ohm and 64 ohm, I then connected the Porter & Davies BC Gigster Drum Stool from our rental inventory to the measurement system, running a plot free-standing, then one with my 220 lb body sitting on the seat. Sitting on it makes a minor smoothing of a 140 Hz resonance as well as a minor change in the 200-300Hz region. Typo occurred when annotating the plot, calling that LF peak 14Hz.

I'm not sure which transducer is used inside this particular seat.

[nickb]

Himm.. I think I'd one of those transducers might be just the thing to make those pesky intermittent faults show up on so much of the gear I see.