Cinema Engineering 7080 graphic valve eq schematic and discussion

[plumsolly]

More exploring and experimenting today and still no luck with the bump. I disconnected the secondaries of the interstage and the bump remained. I checked all the components in that area and they seem fine. It's a little strange that I am seeing it in both my my units. I'd welcome any possible explanations or advice on what to look at.

Thanks,

Ben

 

[emrr]

Anything change if you disconnect C11 or parallel additional C with C18?  Tried disconnecting interstage primary? 

 

[plumsolly]

Hey Doug,

I haven't tried disconnecting C11 - thats what was there (but it has been replaced)

I was thinking about C18 also. I'll try adding another 10 or 20uf, but again that was what was there.

I have disconnected the primary, but it acts a little funny without a load. I think the bump is still there. I will disconnect it again and load it with 150k and check - that way I can take the interstage out of the equation - or not.

Thanks,

Ben

 

[abbey road d enfer]

At the moment, I am trying to get to the bottom of a ~5db 55hz bump which seems to be present on both of my units.

I tend to think the values of R20/C73 and R21/C74 are not correct. Being mica caps, I would think they are 10pF instead of 10nF.
I've done a quick sim and I get the same bump. I question the value of C12, which is generally 1-5uF. Even 0.33uF is not enough. I don't know the exact value of inductance of the primary, so I can only guess, but I would think you need at least 1uF there.

 

[emrr]

True, 0.33 only makes sense with a primary 30k-50k, which seems doubtful.  Try it, and also report pri DCR out of curiosity. 

 

[plumsolly]

Ok. I tried a few things and measured a few things.

DCR of interstage primary is 1.056K

I disconnected the primary and put in 300k to ground, and the bump did go away.

I changed C12 to 4uF and it did not help.

I added another 50uF to C18 and it did not help.

I tend to think the values of R20/C73 and R21/C74 are not correct. Being mica caps, I would think they are 10pF instead of 10nF.

I think you are referring to C13 and C14. They are definitely 10nF. I confirmed this on multiple units.

C12 was 0.25uF originally.

Is it possible that this bump was anticipated/intended and something else was meant to compensate for it and isn't working?

Or the interstages have been damaged by age or dc leakage?

Thanks for your help.

Ben

 

[abbey road d enfer]

I think you are referring to C13 and C14. They are definitely 10nF. I confirmed this on multiple units.

C12 was 0.25uF originally.

Is it possible that this bump was anticipated/intended and something else was meant to compensate for it and isn't working?

I have refined the simulation - still without knowing the inductance values for both xfmrs - and it looks like the R20/C13/R21/C14 values are meant to improve the LF response, by providing a bass-lift starting at 200Hz.
Now we are in a situation where we suspect that the designer(s) have used tricks to compensate the limited performance of their iron; it is thus very difficult to know what their tolerances were. Do you have another unit which would not exhibit this LF hump? or are you just facing the fact that the product works as intended?

 

[plumsolly]

Hey Abbey - both units I have exhibit the same bump. The data sheet specs 20hz to 20k +/- 0.5db. It was a very expensive piece presumably designed by a heavyweight - Art Davis later of Altec fame. So I don't think it's operating as its supposed to right now. A couple things to consider:

I do not have the correct control values connected right now - I have aprox 10k from cathode connections (boost) to ground and aprox 2.6k from plate connections (dip) to ground. As far as I have noticed these values just affect wideband gain, but it is possible that having the correct arrangement here introduces a low end rolloff that my bump is meant to compensate for. I wanted to fix the bump before doing precise tests to figure out these values - could be a catch 22.

C2 is ~2nF, and should cause a high freq boost. Perhaps these are supposed to provide a high shelf that evens out my bump? It seems like a small bypass there should have a large effect because of the nature of the stage. I have these installed, but they don't seem to do much - maybe they are disconnected and I am not seeing it.

Thanks for all your help. I really appreciate you doing the sim.

Best,

Ben

 

[abbey road d enfer]

Hey Abbey - both units I have exhibit the same bump. The data sheet specs 20hz to 20k +/- 0.5db. It was a very expensive piece presumably designed by a heavyweight - Art Davis later of Altec fame.

Not meaning to diss anyone, bright designers are known to use tricks sometimes; when they work, they are called a brilliant inspiration.

I do not have the correct control values connected right now - I have aprox 10k from cathode connections (boost) to ground and aprox 2.6k from plate connections (dip) to ground. As far as I have noticed these values just affect wideband gain, but it is possible that having the correct arrangement here introduces a low end rolloff that my bump is meant to compensate for.

I wouldn't bet on it.

C2 is ~2nF, and should cause a high freq boost. Perhaps these are supposed to provide a high shelf that evens out my bump?

More likely to compensate for the inevitable HF losses due to stray capacitances and transformers leakage inductance.

BTW, I suspect this unit is meant to be terminated with 600r. Have you done the tests accordingly. I know it shouldn't really change the interstage hump, but there is some interaction because of the significant NFB.

 

[plumsolly]

Hey Abbey,

Didn't think you were dissing - Just meant that a 5db bump wasn't likely left there in the final product.

I have kept a terminating resistor accross the output - somewhere in the 600-1k range.

Best,

Ben

 

[abbey road d enfer]

Just meant that a 5db bump wasn't likely left there in the final product.

I tend to agree. I think we can eliminate the fact that the EQ section is not properly terminated and assume the problem is located at the interstage. The only explanation I can think of is that the 0.25uF cap and the primary inductance do not match. Since you've experimented with a much larger cap, it makes me wonder if the inductance value had dropped drastically. In order to verify that, you must find a way to measure the inductance. This has been the subject of a debate there http://www.groupdiy.com/index.php?topic=47515.0
You don't need to measure with great accuracy, since it seems the inductance is way off (if that is the origin of the problem), so you may use one of these unexpensive LCR meters. It may be shorted turns, but it seems extremely unlikely that both units have developed the same defect...

 

[lassoharp]

it makes me wonder if the inductance value had dropped drastically

That's an interesting question.  But wouldn't we expect the resonance bump to move as Ben changed the cap values?  I'm assuming it has stayed in the same spot.

 

[plumsolly]

Here is a plot of C12=0.33uF and C12=4uF.

 

[abbey road d enfer]

Here is a plot of C12=0.33uF and C12=4uF.

This is what I get in the sim with the 74k and 10nF in the NFB! That's what I meant by bass lift starting at 200Hz.
Is this measured at the output or at the interstage?

 

[plumsolly]

Holy Sh*t! I was measuring the bump at the primary. Just switching the generator between 50hz and 500hz, and seeing the change in voltage. I assumed I was looking at the same bump I was seeing on the output. You saying it looked like your sim of the NFB loop made me try shorting C13 and C14. That made a BIG difference on the signal measured at the output. I then switched those caps to .068 and viola - FLAT! (or much closer). All the units I have seen have 10nF in C13 and C14, but mine had a bunch of caps in parallel that were definitely not put in at the factory. What's weird, is I still have the bump at the interstage - So I guess there is a lowend rolloff later (OPT?) that cancels it out. Anyways, I know just have to tidy up those cap values for the best response and I am business. Thanks so much for everyone's help. I have been banging my head against this for weeks.

Best,

Ben

 

[plumsolly]

Just realized shorting over C13 and 14 was a bad idea. I don't think I blew anything up.

Best,

Ben

 

[plumsolly]

Here she is with C12=0.33uF, C13/14=0.068uF.

Comes in at +/- 0.5 as per the data sheet.

I tried some other values in both places. It is a weird balancing act. Raising C12 any higher accentuates that 25hz bump, so the low value must have been intentional. Seems like a weird amalgamation of little adjustments/tricks to get the response flattish.

I will look at those values a little closer and then move on to figuring out the values for the control section.

Thanks again to everyone who stopped by.

Best,

Ben

 

[opacheco]

plumsolly,

Do you found the problem?, do you have a final schematic including your found results?

Thanks
opacheco