V76s build - work in progress

[Ilya]

[WORKING OUT THE CONCEPT...........FINISHED]
[COMPONENTS PURCHASE...................FINISHED]
[MECHANICAL DESIGN............................FINISHED]
[PCB DESIGN................................................FINISHED]
[UNIT ASSEMBLY........................................FINISHED]
[TESTS AND ADJUSTMENTS..................FINISHED]

[FULL DOCUMENTATION WRITEUP - WORK IN PROGRESS]

I will be documenting my TAB V76s preamp build in this thread.
Here's what's given:

• Height - 3U
• Transformers and chokes - Sowter
• LPF - removed
• HPF - uses either real or simulated inductor
• DI option - switched between FET buffer fed into the input transformer or direct input to the first tube grid.
• Input gain switch - split between pad section and resistor feedback network section. This allows for using a single pole switch
• VU meter - buffered from the output transformer primary to keep galvanic isolation intact
• All switching is made via relays
• PSU - DC regulated HT, heaters and aux supplies

The main circuit is straight from the German "Brown book". The only difference is an input gain switch that has 4 poles in the original design. This is a rather awkward configuration that can be split into pad section and gain section. In my version the pad section will be activated via on-off-on toggle switch. This allows to use a single pole switch for input gain.

I'm going to provide for an optional simulated inductor schematic for the HPF. This will allow to use an IC instead of a real 100Hinductor (which is pricey).

My main concern for now is the PSU. I have the regulated HT part designed and it should hopefully work fine. I've thought about the heater section a lot. Unfortunately, it's not possible to put heaters in series and use a 12.6V supply because of different current draw of the tubes. CCS is also not very elegant because of the same problem. My current plan is to use an LDO (probably 1085 type) - with the right input voltage it will dissipate around 6Wt which should be manageable in a large enclosure.

Now, for aux supply I need to have 48V for phantom, +/-18V (or +36V) for ICs and VU buffer, 24V for relays (some of them can be combined to work from 36V) and 12V for DI. I haven't done the exact calculations yet, but I believe that 48V 0.3A should be enough for all my aux needs if I generate all these voltages from a single 48V source. However, since the main power is consumed downstream (+36V and +24V are the main consumers) this leaves me with approximately 12*0.3=3.6Wt of wasted power right of the bat. So I'm leaning towards the +36V being my main aux feed, and getting +48V from a doubler or tripler.

I'll be finalizing the design in the coming weeks and I'm planning to start ordering parts and laying out PCBs shortly.

If any of forum members have any thoughts or ideas, they are welcome.

 

[dmp]

Ambitious!

In my version the pad section will be activated via on-off-on toggle switch. This allows to use a single pole switch for input gain.

Good idea to split them out, especially when doing a DI
I did three pad levels (-48, -36, -24, 0) and calculated resistances to do a 24 step gain switch. Increasing the gain steps over the original is a nice upgrade.

DI option - switched between FET buffer fed into the input transformer or direct input to the first tube grid.

I think FET DIs sound inferior. I would never use a FET option with a tube preamp. A better alternative IMO (if you really want a second option) would be a DI step down transformer (i.e. JT-DB) into the mic transformer.
You cannot do a direct input to the tube grid because if the v76 bias scheme - needs a blocking cap.

I have the regulated HT part designed and it should hopefully work fine. I've thought about the heater section a lot.

Unregulated B+ with AC heaters worked great in my build and the measured performance was very good. (EIN < -120dBu). AC heaters when implemented correctly can match the performance of DC.
If you build the PSU in the 3U case, be prepared for some hum/buzz battles. Sowter offers the chokes in mumetal cans, for a pretty reasonable price increase (I built mu metal shields for the CJ wound chokes AND for the power toroid). External PSU, you may avoid some headaches.

 

[Ilya]

Thanks for your valuable input. Here're my thoughts.

Good idea to split them out, especially when doing a DI

I'm doing this for a client, so there are certain considerations and wishes from him. Currently, we agreed to use switchable -30dB/-20dB pad, and a 4dB per step gain switch.

I think FET DIs sound inferior. I would never use a FET option with a tube preamp.
.........
You cannot do a direct input to the tube grid because if the v76 bias scheme - needs a blocking cap.

Frankly, I'm not inclined to put a step-down transformer before a step-up transformer. I have a nice sounding FET DI block that sounds great with transformer inputs. Besides, adding Jensens for 2 channels is a rather costly option, so I'll stick with switchable FET/grid DI input. Of course, I'll include the blocking cap, thanks for pointing this.

Unregulated B+ with AC heaters worked great in my build and the measured performance was very good. (EIN < -120dBu). AC heaters when implemented correctly can match the performance of DC.

I know that AC heaters can be fine, but I don't want to take the risk. I want to be on a safe side. As for the HT, given the possible variation of more than +/-10% in mains voltage I want DC conditions to be stable and independent of mains voltage. Besides, the pass transistor dissipates only 1Wt of heat, so this is not a big deal.

If you build the PSU in the 3U case, be prepared for some hum/buzz battles

Sure, I have a plan for that. I do have the screened input transformer and the first stage choke. Output transformer and the other choke are not shielded, but they will be tacked into a far corner of the enclosure. I won't be able to make the PSU external, but I do have some shielding material for a toroid as a last resort.

 

[dmp]

Currently, we agreed to use switchable -30dB/-20dB pad, and a 4dB per step gain switch.

You probably know this, but pad positions in original were -13, -23.5, -30 dB
Lowest gain without pad is 34 dB and had steps of 6dB

 

[Ilya]

Yes. I've recalculated pad values for -20dB position, and change the gain steps so that they start from 32dB.

 

[Ilya]

More thoughts on heater PSU.
I'm not feeling very comfortable with loosing 6Wt in the regulator.

Initially, I looked at DC/DC converters, but abandoned the idea because there were no readily available modules that can provide 6.3V output (or they were out of stock with unrealistic lead times). However, after reading some books, I think I found a solution. It is possible to use a through-hole LM2596 switcher IC, and through-hole power inductors are also available (although not a huge choice). The circuit is very simple, and given that I've built quite a few SMPS modules (SMD though) I'm pretty confident that this solution will work fine and will be quiet with additional filtering or cap multiplier. Efficiency is rather good (~80%) and calculated IC power dissipation is 1Wt.
I'll dedicate a separate transformer winding for heaters, so the switching noise will have less chance getting into signal path through this avenue.

 

[Ilya]

VU buffer.

VU meter was requested for this design. I'm not a fan of VU meters and I think they are useless in modern digital recording environment. However, they do have a single advantage - they are great eye candies, albeit expensive.

Anyway. There's a big problem with VU meters - they will introduce distortion if connected unbuffered. And given that I have an output transformer which besides converting impedance provides galvanic isolation, it's not a good solution to slap a VU buffer right on the secondary winding. This will obviously connect one side of it to the circuit ground and the unit looses galvanic isolation feature.

What I'm going to do is take a signal from the primary of the output transformer, feed it into a high-impedance buffer stage and scale down at 9:1 ratio to provide a correct signal level to the buffer IC.

I'm also adding a schottky bridge rectifier and a cap (as suggested by Rod Elliott) for cheaper VU meters that don't have proper bridge and/or correct ballistics.
Since I'm going to use a split ground, input and output are AC coupled, and the circuit is referenced to this virtual ground. I'm also including a CCS for VU LEDs.

 

[Ilya]

Slowly working on the input section. Added a -10dB pad option. Now it's a 3-position on-off-on switch that controls the pad relays. Why? Partly because the front panel layout looks much better with this switch - it sits symmetrically with the HPF frequency switch. Had to make a simple diode OR logic block so that the relay in the "off" position worked correctly.

As a side note, I've measured Sowter chokes. Both of them measured almost 2 times lower than specced. I used DER EE LCR meter, and measurements were taken at 1kHz. 100 Hz made little difference. EQ inductor measured at 90H which is within spec. No idea why I get these lowish measurements with plate chokes. Any thoughts?

 

[dmp]

As a side note, I've measured Sowter chokes. Both of them measured almost 2 times lower than specced.

They should be 400 H and 250 H - what did you measure? And DCR?

 

[Ilya]

They should be 400 H and 250 H - what did you measure? And DCR?

The 400H measures 220H, DCR 12.8K
The 250H measures 150H, DCR 4.5K

I'll try to verify inductance measurements with gen/scope R+L method and see what I get. But my LCR meter gave correct readings, although for smaller inductances.

 

[Murdock]

Inductance varies with level and frequency. When 100Hz gave 220H it could very well be 400H at 40Hz

 

[volker]

The 400H measures 220H, DCR 12.8K
The 250H measures 150H, DCR 4.5K

I'll try to verify inductance measurements with gen/scope R+L method and see what I get. But my LCR meter gave correct readings, although for smaller inductances.

You can't get a correct reading with the inductance meter, because the inductance as per datasheet is specified for a certain dc current, which the meter can't provide.

 

[Ilya]

You can't get a correct reading with the inductance meter, because the inductance as per datasheet is specified for a certain dc current, which the meter can't provide.

Agree, but shouldn’t the inductance drop with DC current applied? I expected to see larger inductance than specified. I’m sure I’m wrong somewhere and Sowter is right. Just want to figure out details.
Also, I can arrange a simple CCS to feed the inductor. How do I measure the inductance with DC applied?

 

[volker]

Also, I can arrange a simple CCS to feed the inductor. How do I measure the inductance with DC applied?

Like you said, use a CCS to feed the dc current to the coil. Through a capacitor that is big enough to form an LC below 20Hz you feed a AC voltage for a range of frequencies. You measure the AC current through the coil to calculate the impedance and then L.
It's important to use a CCS or at least a voltage source with somewhat high output impedance, because otherwise it would load down the signal voltage source.

 

[dmp]

The 400H measures 220H, DCR 12.8K
The 250H measures 150H, DCR 4.5K

Interesting that the DCR is lower than the values from the schematic also (should be 18k & 9.3k)
The plate voltages will depend on this DCR

I measured the inductance of the chokes wound by CJ here and the values were pretty close.

 

[Ilya]

It's probably my meter that gives funny readings on large inductances.
I tried several other methods (including Analog Discovery2 impedance analyzer) and I get vastly different values ranging from 200 to 500H. I'm leaving this for the moment since it's clear that I don't have a reliable measurement solution for now. I hope Sowter did their homework and their inductors are up to spec. Will return to this later, when time allows.

As for the plate voltage, yes it will be different. Should I adjust the cathode resistor to get the correct plate voltage reading?

 

[dmp]

No, I don't think you want to adjust the cathode resistors. Build it to the schematic and see what the plate voltages are.

 

[Ilya]

So what's the strategy then? If the DCR is lower, I'll get higher plate voltage. I can compensate that by a) adding resistance to the choke or b) adjusting the cathode R. Maybe the first solution will be more correct than adjusting Rk.

 

[dmp]

adding resistance to the choke

yes

And it may be fine as is

 

[Ilya]

Mechanical design considerations.

I'm going to split the circuit in several parts. The input section will be mounted on the back panel, besides all the input components its PCB it will also serve as a mount for an output transformer and a choke.
The front controls board will be mounted right behind the front panel on the pressed-in threaded inserts. All the control signals will be delivered via the IDC connectors. All audio connections will be screened.
The input section is almost finished and looks like this: