GroupDIY Bridge compressor.

Help Support GroupDIY:

DaveP

Well-known member
Joined
Nov 8, 2005
Messages
2,824
Location
France
Thought I'd better start off in this new section rather than the Lab.

My Edcor OPT has still not arrived but I managed to rig up a HiFi 8K:15ohms in order to get things moving.  I'm still breadboarding this at the moment.  I find that if I can make it work as a mess then it stands a better chance when its all tidy.

What I'm working on at the moment is separating all the frequency poles.  Although there is no official feedback in the circuit, there is still the possibility of inducing feedback from the high ac voltages on the OPT back to the input stages.  On the original 26W schematic you will notice that the wire taking the CV to the 12AU7 bridge tube is screened, this shows they understood the potential problem.

On the 26W, the high pass poles are 1.4Hz and 28Hz and the low pass is 450kHz or thereabouts.  The low pass due to the OPT is not known but may be of the order of 60kHz.

My circuit replacing the interstage with caps and dif amps is obviously going to add more poles to contend with, so I have to make sure that no two are the same otherwise instead of nice safe 90 degree phase shift we will have 180 degrees and the possibility of ringing and oscillation.

best
DaveP
 

DaveP

Well-known member
Joined
Nov 8, 2005
Messages
2,824
Location
France
The breadboard version of this is working, I've had to make a few small changes to the unbalanced input stage which I'll post later.  There are a few points in the build that will need special attention:-

All wiring around the bridge is quite sensitive to pick up from the CV rectifier circuit and will have to be run in screened cable unless the layout can keep the rectifier and the bridge sufficiently apart.

Using solid state diodes makes the CV ok but they radiate interference over the input stage.  The neat little 6H6 is screened inside its own full metal jacket which is an unexpected bonus.

You can get the amp passing perfect sine waves but as soon as you connect up the diodes to the OPT it all goes to hell, screening and keeping wires short restores the situation.

Using the 6CM6's (6V6 in 9 pin format) triode wired gives about 3.5W output which translates into +35dBm, which should keep most people happy.  Its very difficult to see the start of distortion on the sine wave, as there is just a subtle rounding off of the peaks, which I would guess is mainly 2nd harmonics, which I hope will sound nice if pushed.  The original 6F6's were triode wired too so it should sound very similar but cost less in tubes.

At the moment there is no feedback in the circuit but it is easy to add 3dB with 150k plate to plate resistors from the output tubes to the unused plate outputs on the 12AT7's.  This is made clearer on the attached schematic.  This is not the final version but it shows how the design is moving.

Any comments +ve or -ve appreciated.

best
DaveP
 

Attachments

  • Bridge control amp 14 feb.JPG
    Bridge control amp 14 feb.JPG
    90.4 KB · Views: 504

alexc

Well-known member
Joined
Sep 21, 2004
Messages
2,571
Location
Hobart
Look like you are moving ahead with your design Dave.
Great to see  :)

Could you elaborate some on seperating the poles?

Are you calculating from the design only or is there a technique you use for measuring?
Like measuring gain vs freq at different points in your circuit?

Cheers
 

DaveP

Well-known member
Joined
Nov 8, 2005
Messages
2,824
Location
France
Alex,

The basic formulae are well described in Randolph Aitkin's tech section, but as his site is down at the moment I've attached them for reference.

From the tube data you can get the capacitance data you need and the rest comes from drawing load lines and getting the mu, gm and rp data off the chart as described in Radiotron.

I put all the relevent poles or -3dB points on the schematic so its possible to see if any are the same.  Using the f1 point (see attachment) as an example, by playing with the coupling cap and grid resistors values you can change the -3dB points to get them an octave or two apart.  The idea is to have a step in the high or low roll-off curves so that they don't reinforce each other until the gain is so low it doesn't matter.

So in general, we need grid resistors as high as the tube allows which allows us to use the smallest coupling cap for the chosen -3dB point.  I'm sure the tutorials in my references can explain it better than I have.  The dif amps I've used and the concertina on the first stage have the added advantage of multiplying the value of the grid resistor so very small coupling caps can be used whilst still retaining the frequency response.

All of this is much more relevent to feedback amplifiers but feedback can creep in through the power supply and bad layout, so I thought it best to design the problems out just in case, if this compressor is made by guys with varying layouts and expertise, then hopefully it won't matter too much if they make a few bad choices, in other words I'm aiming for a forgiving circuit.

best

DaveP
 

Attachments

  • Circuits and Equations.doc
    25 KB · Views: 118

alexc

Well-known member
Joined
Sep 21, 2004
Messages
2,571
Location
Hobart
Thank you - that certainly refreshes the memory.

I'd done this sort of thing many many years ago with solid state but never applied it to feedback tube gain stages.

Seeing as I have a few projects with feedback, I'd better do some more reading.

Cheers
 

DaveP

Well-known member
Joined
Nov 8, 2005
Messages
2,824
Location
France
Thought I'd better update on progress so far.

After wrestling with the oscillation on my 26W copy due to excess positive CV on the grids, I felt like a break.

I've breadboarded a 26C and it appears to work without problems.  I got almost 14dB GR which is on spec.

The interstage is critical to this project and you need two!  I decided to use the VT1399 I didn't use in the BA-6A last year so I took a chance and ordered another.

While I'm waiting for that to arrive I tested the design with two 30V mains transformers.  Obviously their frequency response is hopeless but for a 1kHz test they are fine.

I tested  the inductance on each leg of the VT1399 by measuring the resonant frequency with 5 different value capacitors and they worked out at 6.6H each.  They will therefore only give good bass response with a low source resistance.  f-3dB = Rout/ 2pi x L (Henries).  I got around that by fitting a C/F to the stock tube with an Rout of 220 ohms, this gives 5.3Hz which should be ok.

To avoid using yet another Interstage for the phase splitter I used the other half of the C/F tube as a cathodyne, a 6BQ7-A works a treat for this duty.  I won't put the circuit on line yet as its still work in progress.

Edcor have still not shipped my OPT (ordered Jan 18th) but they intend to ship on March 5th.

Thats about it for the moment, it looks like its going to be 2 weeks before I have all the parts and can start planning the metalwork.

best
DaveP
 

DaveP

Well-known member
Joined
Nov 8, 2005
Messages
2,824
Location
France
Thanks guys,

I've virtually finished the design work today and just need some details tidying up in the lab next week, then I'll post the schematics.  Its a pity the transformers take so long to arrive, I can't finalise the chassis shape until I get them.

The nice feature of the 26C is that the bridge tube is just used as a variable resistance, it doesn't amplify any signal.  It gradually disconnects a short to the windings so that they go from being in phase to out of phase, this is what causes the GR, very slick.  So should be thump proof.

From my calculations, it doesn't look like the 26C will be capable of 0.1mS attack but 1mS should be possible, more next week.

best
DaveP
 

DaveP

Well-known member
Joined
Nov 8, 2005
Messages
2,824
Location
France
Gemini,

No I don't think so, it has to have at least 3 separate windings/tX so that they can be wired in antiphase.  When the tube is at 0V normally, it shorts out the antiphase connection, but when the CV goes -ve the rp rises and disconnects the short so that the TX goes part antiphase and starts GR.  The VT1399 has 4 separate windings.

Hope that makes sense!

best
DaveP
 

gemini86

Well-known member
Joined
Sep 22, 2008
Messages
2,476
Location
Eugene, OR USA
No, that makes sense...

If we have enough interest, I'm pretty sure they'll do a custom job. They're usually about the same price as "stock" units, since they don't actually stock most things.
 

DaveP

Well-known member
Joined
Nov 8, 2005
Messages
2,824
Location
France
Update.

The second VT1399 interstage arrived from "Livinginthepast" and the good news is that they work.  Feeding the TX from a C/F is fine but the choice of coupling cap has to be chosen with care to avoid unwanted resonances at the bottom of the audio range.  There is a problem with the top end at the moment because of a steep rise in the high freq response which will have to be tamed.
Any ideas on that are welcome.

I had to find a cheap substitute for the 6C8G used in the bridge.  It needs a low resistance when fully on and the highest possible when it goes negative, I initially tried a 6N7 and got 14dB GR but that uses lots of heater power so I tried a 6DT8.  I had not realised that this was basically a 6v version of a 12AT7 by GE, it is an excellent tube at $3.00.  It has a 100V max on the heater/cathode limit instead of 90V and a very high resistance when the grid goes way negative, using this tube I got the GR up to 16dB.

all for now.
DaveP

 

emrr

Well-known member
Joined
Apr 12, 2006
Messages
7,923
Location
NC, USA
Very good.  Guessing you have a top end resonant rise from the low Z drive? 
 

lassoharp

Well-known member
Joined
Jan 3, 2009
Messages
2,067
Location
USA
Sounds great!



cheap substitute for the 6C8G used in the bridge

Possibly 6SN7/12Au7/6cg7   

or pair of 6C4s

I'm looking at two general purpose lo mu triodes, so the list would be even longer.





audiophools have ruined the market on 6C8G.  :eek:
 

DaveP

Well-known member
Joined
Nov 8, 2005
Messages
2,824
Location
France
Doug,

I seem to have cured the problem with a zobel between the two interstages, when I looked at the scope with a square wave there was too much ringing, big clue!

Alan,

The mu of the 6C8G  is about 30.  The 6SN7 is way too expensive for a cheapskate like me.

I've made an alternative CV rectifier from another 6DT8, common cathode direct to cathodyne to give a balanced supply for two 1N4007's; output can easily make the -10V needed to switch the bridge.  I did not use the original 6R7 because it needs yet another interstage and I'm running short of room.

best
Dave

 

tommypiper

Well-known member
Joined
Jun 5, 2004
Messages
1,281
Location
bay area
DaveP said:
Doug,

I The 6SN7 is way too expensive for a cheapskate like me.

Dave

I wouldn't rule it out on that basis.  There are some good sounding cheap Chinese versions.  ($10 each, with free int'l shipping.)  A group of friends tested them against other NOS, and the Chinese sounded best.  We were all surprised.  I heard them with my own ears during the test...  FWIW.
 

DaveP

Well-known member
Joined
Nov 8, 2005
Messages
2,824
Location
France
Tommy,

I have nothing against Chinese tubes although the ones you found are 6SL7's not 6SN7's.  As long as the triple oxide used for the cathode emission is pure enough, and the metal doesn't outgas, the vacuum is good and the structure is not microphonic, it should make no difference what nationality makes them.

This project has as one of its objectives, to use neglected tubes because certain types are being touted as "must haves" and I believe it is all hype to separate people from their cash.  This design is not tube dependent, so the DIYer is free to sub in other types depending on what's available, including Chinese types.

Edcor have advised me my OPT has shipped so that should arrive soon as long as UK customs send me a card advising me of the duty to pay.  It is not beyond them to just sit on it and wait for me to contact them.

best
DaveP
 

skal1

Well-known member
Joined
Jun 4, 2008
Messages
1,205
Location
Birmingham,uk
HI Dave p


what is the price of the custom made o/p tx from edcor, glad to see the proto is chuging along , if you need a beta tester i am ya man


peace
skal1
 

DaveP

Well-known member
Joined
Nov 8, 2005
Messages
2,824
Location
France
Skal,

The Edcor 8k:600 is from their published range at $39, then there is $40 for shipping plus import duty and VAT, not quite so cheap after all!

I'd be happy for you to test it before I ship it stateside, what's your set-up?

best
DaveP
 
Top