Tested 5 VR Tubes In Eros

[Grainger49]

I received 4 12BH7As yesterday.  I wondered how much variance the tubes would make in the voltage readings on the boards.  I measured the two C4S boards voltages with the stock EH tube and the 3 RCA, 1 Sylvania I just got.  Input voltage is 124.5V out of the wall.

The 1V on Kreg ranged from 1.596V to 1.717V.  This was the greatest variance and I'm betting it has to do with the capacitors.

OB is designed at 100V, it ranged from 95.1V to 97.8V.

Breg designed at 100V ranged from 96.5V to 98.7V.

IA is designed at 225V and ranged from 219.6V to 223.7V.  Interestingly all of the left PCB voltages on this pad were 219.XV and the right PCB voltages were 223.XV.

OA is designed at 170V and ranged from 157.1V to 158.3V.

Nothing except Kreg was off by a substantial percentage.  And the only way to look at is is percentage not volts.

I'm just throwing some data out there.  If I had rolled the audio signal tubes the voltages might have changed a little bit.  But what I was interested in was the regulated voltages

 

[grufti]

OA is at the low end of good at 157.1 to 158.3. You want 60+V difference between Breg and OA.

 

[Grainger49]

The EH was new this fall.  The other tubes were advertized as NOS but only one looked like it had been in the box since new.  The voltages didn't stray much at all.

I'll look at the specifics and see if I got the 60V between Breg and OA on each tube.

 

[Grainger49]

Looking at just OA and Breg for each tube:

Design:         OA 170V       Breg 100V   Delta 70V
EH LPCB         158.3V            96.5V      61.8V
EH RPCB         157.5V            98.3V      59.2V
RCA 1 LPCB     157.2V            97.8V      59.4V
RCA 1 RPCB     157.6V            98.4V      59.2V
RCA 2 LPCB      157V            97.4V      59.6V
RCA 2 RPCB    157.1V            98.7V      58.4V
RCA 3 LPCB    158.1V            96.6V      61.5V
RCA 3 RPCB    157.3V            98.6V      58.7V
SYL LPCB        158.3            96.4        61.9v
SYL RPCB       157.6V            98.4V      59.2V

Everything lined up as best as my old eyes can...

As you can see of the 5 tubes not a single one has 60V or more on both regulator boards.

So what am I doing wrong?  What component will bring these closer to design parameters.  I assume it is a resistor on the C4S boards, right?

 

[Paul Joppa]

These are plenty close enough.

The actual parameter that matters is the bias on the 6922 - the voltage between grid and cathode. It should be a bit over 1.5 volts; this assures that there is no grid current, and hence that the grid impedance is high enough that it will not affect the equalization, nor will it contribute to distortion.

As the 6922 ages, its emission will decline, and the plate voltage will rise. It should be close to 70v, averaged over the lifetime of the tube - which is how the manufacturer's specifications were derived. The 70 volts is the same as the design plate voltage in a Seduction, where it is pretty common to find 55-60v with fresh tubes. (In Seduction you have the LED to assure the correct bias.) In Seduction, when the plate voltage rises to 90v the tube is worn out. In Eros, I'd check the bias voltage, and replace the tube when it drops below 1.2 volts - but first, I'd listen to a new tube, and keep the old one if it still sounds good.

I choose design parameters like voltage based on the specs of the tube, which are averages of many tubes as tested by the manufacturers - mostly many decades ago, of course. In the manual, we usually have voltages as measured on the prototype, with the particular tube we had at hand when the manual was written. But individual tubes do vary; that's why we specify an acceptable range.

 

[Grainger49]

Paul,

Thanks for the design voltages and the reassurances.  I didn't imagine that 5 VR tubes could all be wrong.

 

[Noskipallwd]

Grainger, just wondering if you heard any difference in SQ between the different VR tubes.

Cheers,
Shawn

 

[Grainger49]

Shawn,

I did all the measurements on my work bench.  And the Eros is back there because I lost the left channel.  I verified it was the Eros by swapping input and output cables.  The Left speaker came alive with each swap.

This was caused by my starting to roll output caps.  I have the stock Solens pulled out and jumpers to the front of the Eros so I can check the Solens vs. KK Teflons vs. Mundorf Silver/Gold/Oil.

I know it is self inflicted and will be easy to find when things calm down.  (It is fixed and the Mundorf are soldered into the circuit)

 

[ALEXZ]

Grainger,
Maybe I'm missing something, but why are you expecting changes in the EF86/6922 bios ?
You are changing regulator tube, that may move 225v output for a volt or two up or down. 
Let say HT changes 3 volts, 6922 is 100v above ground, that  lives 125 volt  for 6922 and current source.
6922 plates is about in the middle of 125v. comparing 225v HT with 62v on the plate of 6922, rough estimate that 3v change in HT brings less the 1v change  for 6922 plate voltage. All that is just a rough estimate, without any considerations  for direct connected EF86, DC feedback etc. just to demonstrate how little mpact may VR tube have on Ef86/6922

 

[Grainger49]

It seems to have changed the regulated voltage and the incoming voltage from the regulator tube.

My Eros is on a voltage regulated supply, not the voltage from the wall.  So incoming AC voltage is not a variable. 

I was just curious what the 5 different tubes would change.  A solid state regulator would give you the same output voltages pretty reliably.  Because of the internal differences in tubes I was curious what the range would be.

Since all measurements were taken on the bench I have no idea if any of the tubes sounded differently.  The stock EH is still in there.

 

[ALEXZ]

It seems to have changed the regulated voltage and the incoming voltage from the regulator tube.

What's "incoming voltage from the regulated tube" ?

 

[Grainger49]

IA reflects the regulated incoming high voltage.

In the first post it ranges from 219.6V to 223.7V.  Pretty much spot on.

 

[ALEXZ]

Correct me  please , if I wrong, but there is no

 

[grufti]

That's how I hope it works, because that is how I imagine it works. Let's hope PJ doesn't disagree.

Correct me  please , if I wrong, but there is no

 

[Grainger49]

The schematic shows 275V DC unregulated coming into the C4S board feeding the VR tube.  Each one shows 225V DC as an "Output."  The first goes to socket B C4S PC board pad IA, the second goes to socket A C4S PC board pad IA.

I have to take it that the shunt regulated output feeds the IA for each of the EF86 C4S boards.

Am I missing something?

 

[Doc B]

You guys are hanging up on semantics. Grainger is talking about the raw B+ feeding the load on top of the reg vs. the regulated voltage, which is measured at the bottom of that load, the top of the shunt reg and the top of the loads on the voltage amps. The shunt reg isn't gonna work without a load, and so input and output voltage could be an appropriate description relative to the top active load.

 

[Grainger49]

Amen!  It is a Moot point.