670 signal amp/control amp woes

[skal1]

Hi Dave p  ,

been playing with this all day so ,i have changed wire and feedback pots a million time or so, just rechecked the side chain output tx primary again  got 165vac across primary ..


cheers

skal

 

[abbey road d enfer]

Hi Dave p  ,

been playing with this all day so ,i have changed wire and feedback pots a million time or so, just rechecked the side chain output tx primary again  got 165vac across primary ..


cheers

skal

That's 82.5V at the plate for 1.8 at the grid. Seems too high. Is it with or without load at the secondary?

 

[skal1]

this is with feedback in place abbey road d , and loaded  with bridge and tc network.

by the way, i found a problem with the the ratio switch , i have corrected that now i think this was part of the problem.

I think i will have to adjust the the fix grid bias again and change the grid leak resistors, does anyone now what gain the pentode should be delivering to the bridge.


cheers

skal

 

[abbey road d enfer]

this is with feedback in place abbey road d , and loaded  with bridge and tc network.

by the way, i found a problem with the the ratio switch , i have corrected that now i think this was part of the problem.

I think i will have to adjust the the fix grid bias again and change the grid leak resistors, does anyone now what gain the pentode should be delivering to the bridge.


cheers

skal

My simulation is not too far from your practical results. Using 5k as primary impedance (according to Hammond specs), I get about 60Vpk at the plate for 1.8Vpk at the grid. But with the EL84-correct 8k, I get the same value as you. That means that the secondary of the xfmr is probably light-loaded.
This seems to indicate that the reflected load is higher than

 

[skal1]

Did a quick test , i took the bridge off and measure across 70 tap i get 50vacrms this is with a 0.700mv  at the grids of el84.

then i reconnected bridge and measure across 70 tap again and  got 11.55vacrms this is with a 0.325mv  at the grids of el84.

it looks like something is loading down the txt hard , what could it be , bridge or the cv line to the grid.

skal

 

[abbey road d enfer]

Did a quick test , i took the bridge off and measure across 70 tap i get 50vacrms this is with a 0.700mv  at the grids of el84.

then i reconnected bridge and measure across 70 tap again and  got 11.55vacrms this is with a 0.325mv  at the grids of el84.

it looks like something is loading down the txt hard , what could it be , bridge or the cv line to the grid.

skal

I guess you got your units wrong. 700mV is 0.7V and 325mV is 0.325V.

If I read you right, the pentode gain is ca.70 without the bridge and 35 with the bridge. So the bridge loads the pentode stage by about 6dB, which is somewhat surprizing. Indeed, the steady-state load is minimal; the heavy stuff is charging the timing capacitor, but once it's done (takes a few millseconds) the equivalent load is about 1/10th of the nominal load there (I estimate the nominal load at about 200 ohms).
I must start investigating the old-fashioned way: elimination.
First, disconnect the CVline from the input transformer's center-tap and groun dthe center-tap, so the whole chain should have stable gain.
Then start by connecting just the bridge, then the timing caps, and check the voltage at the output of the xfmr.

 

[skal1]

ok ,

{No bridge}  got  957mv across grids of el*4's  at the secondary 70v tap getting 52vac rms

{bridge 3900 connected ,tc disconnected }  got  1.548v across grids of el*4's  ,and at the secondary 70v tap, across bridge getting 50.9vac rms

{bridge 3900 connected ,tc connected }  got  1.600v across grids of el*4's  and at the secondary 70v tap, across bridge getting 50.5vac rms

{bridge 3900 connected ,tc connected and cv to connected to ct }  got  396mv across grids of el*4's  and at the secondary 70v tap ,across bridge getting 11.3vac rms

hum looks to me like the amp has a heavy load , looks to me like it is the input tx 1:4  step , correct me if i am wrong.

cheers

skal

 

[DaveP]

That amp was designed to have an output impedance of 100 ohms so something is wrong somewhere to drag the output down so much.  Have you checked any resistors and caps in the CV circuit?  I have mis-identified resistors before now where the colour coding has been poor, are the caps holding charge or shorted out?

DaveP

 

[skal1]

Hi davep,

i am only using the first  position  220k||68k||2.2uf

resistor || = 52k

cheers

skal

 

[abbey road d enfer]

hum looks to me like the amp has a heavy load , looks to me like it is the input tx 1:4  step , correct me if i am wrong.

How could an input xfmr load the output of the side-chain amp?
What I see is that as sson as you connect te bridge AND the 3900r, the gain drops by about 3-4dB.
Are you sure the bridge is right? Are you sure the 3900r actuallys measures 3900 ohms?

 

[skal1]

well haven't measure 3900k for a while might check that now and get back

 

[skal1]

just measure the 3.900k coming in at  4k  and bridge is gone,  just put in a 6AMp  bridge  ,going to do a bit of testing now, to see what happen...

cheers

skal

 

[skal1]

ok back on it again , messing with the signal amp right now .

i can afford to buy one of the sowter transformers for this project but i can decide which 1 to go for, at the moment  i am using 1:4  600:10K on the input , 4:1 10k:600  on output  and 1:4  600:10K for the side chain input , my ? is , on which part should i spend the money.

regards

Skal

 

[abbey road d enfer]

ok back on it again , messing with the signal amp right now .

i can afford to buy one of the sowter transformers for this project but i can decide which 1 to go for, at the moment  i am using 1:4  600:10K on the input , 4:1 10k:600  on output  and 1:4  600:10K for the side chain input , my ? is , on which part should i spend the money.

regards

Skal

When you write "signal amp", I guess it's the one around the transconductance gain cell.
I would say that T102 is the most critical.
T101 is fairly standard and being driven by a low-Z source should give correct performance, but T102is the one where intrinsic performance is the most important since it is driven by a variable but rather high impedance.

 

[skal1]

ok back on it again , messing with the signal amp right now .

i can afford to buy one of the sowter transformers for this project but i can decide which 1 to go for, at the moment  i am using 1:4  600:10K on the input , 4:1 10k:600  on output  and 1:4  600:10K for the side chain input , my ? is , on which part should i spend the money.

regards

Skal

When you write "signal amp", I guess it's the one around the transconductance gain cell.
I would say that T102 is the most critical.

Yes  ,  thats the output tx  , could i also use it in front end of transconductance gain cell just to see what voltage i need thier.

regards

Skal

 

[abbey road d enfer]

ok back on it again , messing with the signal amp right now .

i can afford to buy one of the sowter transformers for this project but i can decide which 1 to go for, at the moment  i am using 1:4  600:10K on the input , 4:1 10k:600  on output  and 1:4  600:10K for the side chain input , my ? is , on which part should i spend the money.

regards

Skal

When you write "signal amp", I guess it's the one around the transconductance gain cell.
I would say that T102 is the most critical.

Yes  ,  thats the output tx  , could i also use it in front end of transconductance gain cell just to see what voltage i need thier.

regards

Skal

You could use it for the specific purpose of evaluating voltage, but the optimization of an input xfmr and an output xfmr are quite different.

 

[skal1]

ok back on it again , messing with the signal amp right now .

i can afford to buy one of the sowter transformers for this project but i can decide which 1 to go for, at the moment  i am using 1:4  600:10K on the input , 4:1 10k:600  on output  and 1:4  600:10K for the side chain input , my ? is , on which part should i spend the money.

regards

Skal

When you write "signal amp", I guess it's the one around the transconductance gain cell.
I would say that T102 is the most critical.

Yes  ,  thats the output tx  , could i also use it in front end of transconductance gain cell just to see what voltage i need thier.

regards

Skal

You could use it for the specific purpose of evaluating voltage, but the optimization of an input xfmr and an output xfmr are quite different.

i will have to look at optimization of the input xfmr later.

cheers

skal

 

[DaveP]

I agree with Abbey,

The OPT for the signal amp is the most important TX and I don't think you can't get it anywhere else other than Sowters.

Work out on paper the levels going through each stage of the compressor to see if they make sense.  You must have that small pad between the Signal amp output and the control amp input for isolation, but if your levels don't quite match up you could tweak the pad a little to sort that out.  If you don't get the levels in the right ballpark, then your compression curves could end up in  completely the wrong place.  That is to say they could kick in too early or too late.

DaveP

 

[skal1]

Ok this what i have set up for the GR amp.

Signal gen out is .245 mVrms to primary of 600:10k tx , voltage across secondary is 1.016 Vrms  ,this  is injected into the grids of the GR amp which has a set down Ratio of 4:1 ,the  voltage across the  primary is 24 Vrms and the secondary voltage is 6 Vrms , this is without a load across secondary of transformer.

How would one calculate the db gain of the GR amp, please let me know if i am using the wrong terminology for the voltage , don't want anybody getting confused.
 

regards

skal

 

[abbey road d enfer]

Ok this what i have set up for the GR amp.

Signal gen out is .245 mVrms to primary of 600:10k tx , voltage across secondary is 1.016 Vrms  ,this  is injected into the grids of the GR amp which has a set down Ratio of 4:1 ,the  voltage across the  primary is 24 Vrms and the secondary voltage is 6 Vrms , this is without a load across secondary of transformer.

How would one calculate the db gain of the GR amp, please let me know if i am using the wrong terminology for the voltage , don't want anybody getting confused.
 

regards

skal

Quite easy, since we know the input and output levels:
I guess you meant 0.245V or 245mV for the input level, so the actual overall gain is 6/0.245 or 24.48 or 27.8dB.
Since the input and output xfmrs have the same ratio (in reverse), the actual grid-to-plate gain is the same.