Zach's 700B Thread

[Gepetto]

Q7 and Q10 are always on Mark.

Q10 derives its collector and emitter current from the B+ rail through R23 as a static voltage source.

R34 is the primary equivalent of R23 for Q7's emitter and collector current which basically scavenges this current off of the bias current produced by the upper half of the amp, Q11 and R38 at idle and off the emitter currents of Q13-Q17 when the amp is actively driving the upper half of the signal. The source for Q7's emitter and collector current is thus dynamic, not static and thus unlike the upper half.

If the global feedback demands of the amp are such that Q5 is calling for more "down" signal and Q7 cannot provide that through R34 alone, the voltage drop produced by the current flowing through R34 will eventually turn on D13 and provide "turbo boost" allowing more current to flow through R35 as well. In the case of large reactive loads on the output this is often important in order to "tame" that load.

This is the readers digest summary of how this works.

The local feedback works a little different. In the presence of current drive from Q7 which is a classic emitter follower connected to the output signal of the amp, current amplification of Q7s current is provided by Q12, Q14, 16, etc. UP TO THE POINT that Q12, Q14, 16 etc. saturate to the point that they begin to cut off the emitter of Q7. When that point is reached, the current drive provided by the collector of Q7 to Q12, Q14, 16 etc. dries up and an equilibrium is reached and it drives no farther than demanded by the emitter node of Q7. This is the local feedback path I am referring to. The self regulating loop formed by Q7, Q12, Q14, etc.

There is significant voltage as well as current gain provided by Q12, Q14, 16 etc which can lead to instability in this local loop. This is unlike the upper half which is a straight emitter follower triad (Darlington) configuration with NO local feedback.

Tricky to describe, I hope I did it some justice.

 

[mlucitt]

Joe, Nice description. This perfectly explains the lack of symmetry in the output architecture of the PL main board. Local FB is always more complicated for me to understand than global FB. You mentioned the PNP transistors are more expensive and this is still true PNP 2N5416 = $2.23 compared to the equivalent NPN 2N3439 = $1.83 each at Mouser.

Mark

 

[Gepetto]

Yes on the PNP's Mark, still more expensive and less prevalent.

When you work in power MOSFETs (which I do all the time), the same is true. N channel MOSFETs are jellybean, while P Channel MOSFETs are more expensive, much smaller selection and not as capable.

 

[laatsch55]

Hi Lee
Need to know what you mean a little more before answering?

Do you mean why did they use the quasi complementary structure on the bottom half?

Guessing what you want to know, PL used all NPN devices because back in the day this amp was created, NPN transistors of this class were all that existed. PNP power transistors were rarer and pricier(still are today). NPN's are easier to fab in silicon than PNPs and are lower cost. PNPs are easier to fab in Germanium (and lower cost) than NPNs.

Did I get at the heart of your question?

Yes you did...

 

[Zach C.]

Hi fellas,

I just wanted to let you know that I am coming back. I got sidetracked with some school stuff, so it'll be a few days to a week before I can do much more on this project unless something changes.

Thanks, for all the help!

Zach

 

[wattsabundant]

Zach and everyone else,

I haven't checked in here for a while. I've seen this problem and just fixed it last week on one of my own 700b's. It's driving me up the wall cause it defies all logic. The solution is to ground the input jack to the output binding post as the amp was originally wired. I was thinking I had a test equipment ground loop and used isolation transformers to no avail. I've tried evey conceivable combination and this is all that has worked. see the attached picture. I'm open to comments as to the origin of the problem and the solution.

 

[laatsch55]

Don, It's happening on the 400 boards too. Could not get the left channel to settle down, looked like an oscillation. I think it has to do with the outputs being grounded when the relay drops out. take a look at it Don, somethiong there isn't right.....
Anyway , pulleds the relay out and even without a zoebel, steady as a rock.

 

[laatsch55]

Have you tried cutting the trace from the NC contact to the speaker ground post??

 

[wattsabundant]

First a comment about the relay. There are no relays rated to break 100VDC that will fit in the chassis. Regardless what is used, the contacts will weld under a fault condition due to the arcing. There is no snubber that can fix it. I know from experience. During development I intentionally sent 100VDC through various relays and watched it on a scope. They all welded. I tried all of my bag of tricks. So to protect the speakers when the relay drops out it shorts across the speaker. Anything that defeats this defeats the protection.

Last night I verified the 400 problem that Lee mentioned. Amp worked fine until I installed the relay. Oscillated once installed. Tried all kinds of grounding techniques. No good. Looked at the zobel as it is the only difference. Caps on old network and new ones are all within .002uf of .1uf. Changed the resistor from 4.7 to 10 ohms. Amp works great. used two parallel 15 ohms (=7.5 ohms). Amp works great. Measured about 15 old resistors (2 x 10 ohm in parallel = 5 ohm). All measured above 7 ohms.

When I first developed the board I used 5.1 ohms. Then for some reason they weren't available and I switched to 4.7 ohms. I will be looking at a new value north of 5 ohms. In the mean time 7.5 ohms is very stable under all conditions and is not affected by grounding technique. Will investigate further when time allows. Day job is going to get int he way for a couple weeks.

For interesting reading on snubbers, google paktron quencharc. Even they don't work on 100VDC.

 

[Zach C.]

...I've seen this problem and just fixed it last week on one of my own 700b's...

This is an interesting development. When you said that you tried everything, I assume that to include the "star" grounding approach proposed in this thread *. I interpret the advise I got here as meaning, "it's a needed improvement, so do that first and then address your other problem." Did you try this also? I don't want to reconfigure my ground set up only to switch it back.

Also, are we sure that all three of us are chasing the same problem? Mine does not settle when the only change is to lift the output zobel out of circuit. Whereas laatch says his is fine w/o the zobel. Mine draws much less current (dimmer bulb) with just the zobel lifted than it does with the DC board in place, but it still won't settle.

At the moment, I'm still of the opinion that my problems lie somewhere other than Don's board.

I think I may sit back for a few days and see what happens after some exchanges amongst the pros. If you guys had any idea how long I've been dinking around with this (off and on), you'd know how little hurry I'm in. Sadly, I don't think I'll have much to contribute to the discussion. I'll have a lot more free time in a couple of days, so I could try some stuff if needed/ thought potentially helpful.

Thanks,

Zach

* as posted by mlucitt
http://forums.phxaudiotape.com/attachment.php?attachmentid=7781&d=1360981853

 

[laatsch55]

Zach, the 400 I'm having problems with did not have Don's zoebel network on it. I had substituted a stock PL zoebel (2 10 ohm 3watt resistors in parallel ,seriesed with a .1uf poly cap). That was the configuration I was having problems with. I bypassed the board, but also put a zoebel in place of the board. The 700 board is a whole different animal than the 400 board, but i have just put Gibsonian's 700 Ser I back on the bench. Some of my problems with it may be related to this board anomaly.

 

[laatsch55]

Zach, I have installed a lot of Don's boards. Have not had any problems related to them except now. With the stock zoebel in place is your amp stable??

 

[Zach C.]

Zach, I have installed a lot of Don's boards. Have not had any problems related to them except now. With the stock zoebel in place is your amp stable??

Yes, in the stock configuration, I get good clean sine waves and a solid 350W RMS per into 8r. From there, if I simply lift the output zobel things get funky. Amp will not settle, and maintains a dim bulb.

If I actually try to install Don's board with the recommended wiring configuration, the amp draws massive (scary) current. I have not tried the ground mod that Don mentioned yesterday- i.e. grounding the speaker output and input jacks together with his board in place.

It probably goes without saying, but I am using jumpers to put things back to factory. It is then very easy to try other things.

In your previous post, I initially thought you and I were discussing different things, but after re-reading it, I think we are talking about the same thing, but looking at it differently.

I don't think Don's board is the problem here, but I think that attempting to install Don's board has caused an otherwise unknown issue to come to light. But, what the hell do I know?

Zach

 

[laatsch55]

Not necessarily, there may be an issue with Don's board. In the stock configuration , if everything is fine, and with Don's board it isn't, then there is a problem with the board. There are no sacred cows here. wede try to be corrrect in our assessments of the problem however. i have been known to espouse on an issue only to have Joe come back and tell me further investigation would have led to a different conclusion. but in this case where Don has alerted everyone there is an issue, he's fair game!!

Also, now that the stock config is fine, you can throw Joe's board in with no fears....

 

[laatsch55]

If I actually try to install Don's board with the recommended wiring configuration, the amp draws massive (scary) current. I have not tried the ground mod that Don mentioned yesterday- i.e. grounding the speaker output and input jacks together with his board in place.



There have been some very good minds involved in eliminating this particular connection, it causes a lot of noise, and the process involved in it's elimination has been proven.

 

[Zach C.]

Not necessarily, there may be an issue with Don's board. In the stock configuration , if everything is fine, and with Don's board it isn't, then there is a problem with the board. There are no sacred cows here. wede try to be corrrect in our assessments of the problem however. i have been known to espouse on an issue only to have Joe come back and tell me further investigation would have led to a different conclusion. but in this case where Don has alerted everyone there is an issue, he's fair game!!

Agreed, and I understand that Don's board may be causing further issues. As I understand it though, the amp should settle and run without the factory zobel in place, since it is meant to make inductive speaker loads "look" resistive to the amp to prevent oscillation. No inductive load would then mean no need for the zobel. I may well be confused about this though. Would love to hear more on this point.

Also, good to know about the cow situation- that's the only way to get to a real answer.

Also, now that the stock config is fine, you can throw Joe's board in with no fears....

Mine has been in this state since I came to this board i.e. working in the stock config. Other than the broken q4 that I found while talking about this with you guys, the amps state is as it was when we started. Again, I am CERTAIN that q4 was not broken during my previous repair attempts. It turned out that I had those q's in stock, so the amp is back to it's initial state. DC offset is a bit lower now (under 10mV), but otherwise it's as before. I also got the a/c rewire done, but not the star ground...yet.

That scares me a bit due to the situation I mentioned above, and it's why I came here to find out first. I don't want to blow up the new board, or need to switch back to the original solve a problem. Also, it's just nice to know why there is/ was a problem.

thanks,

Zach

 

[Zach C.]

There have been some very good minds involved in eliminating this particular connection, it causes a lot of noise, and the process involved in it's elimination has been proven.

I know that star grounds are the best approach, and that's why I asked Don for more details about what he tried. If I do the star ground thing, and then turn around and install his board in the way he recently suggested, then I really haven't done anything other than possibly screw up a workaround that Don found to what I suspect is a deeper problem with the amp and is itself a kludge, hence his apparent dissatisfaction with it. I'd just be tying them right back together.

I hope that makes sense.

Also, I don't mind dinking around with this a bit on my end (obviously), but would prefer to avoid things Don already tried, or that just won't lead anywhere.

Zach

 

[laatsch55]

Doesyour 700 havew the bypass caps on the 3 terminal strip on the back wall?? If not it could get funky without a zoebel. When you lifted the zoebel in stock mode did you have the amp loaded? What was your input and ouput voltages?

 

[kevin]

MAN... This is one hell of a serious thread......I cant wait to see what fixes this thing.... You guys are just unreal SMART !!!!!! LOL... I'm going to say this again....... It's such a privlage just to be on here with you guys !!!!!!!! WOW !!!!

 

[wattsabundant]

As I recall Zach bought the board from me a year ago and he has been fighting this since then. I'd forgotten all about it. Early on in this thread I thought I saw where he was using a scope and saw an oscillation. Maybe that was somebody else. The 700B fix I previously mentioned returns the amp to the stock, as built, configuration which ties the input ground to the output binding post. Zach, If you want me to look at your amp send me a P.M. I'd like to put this to bed as much as you would.