V1 & V2 Plate Voltages - A critical observation

['67_Plexi]

In over 20 years of building tube amps I've never come across an amp where the the V1 & V2 plate voltages are so critical. It appears that there is really only a very small window of acceptable tone, this falls in line with Dogears insistance of aiming for the 200V mark. He is totally correct, go over voltage and the amp is harsh, go under and it's dull.

The point I'm getting to is the fact that the preamp tubes you use play a big factor in this. The 'book' test current draw for a 12AX7 is 1.2mA per triode.
I use all Mullard tubes and they are normally bang on the rating. However modern tubes are rarely even close. This change in current draw can throw the plate voltages on V1 and V2 way off. It is important you tweak the power supply dropping resistors using the exact preamp tubes you intend to use and not diverting from them.

I compared the following new 12AX7's and recorded the V1a plate values:

1966 NOS Mullard (reference) - 201V
New Production Tung-Sol - 211V
Groove Tubes 12AX7R - 212V
Groove Tubes 12AX7M (Mullard Clone) - 205V
70's NOS RCA 12AX7A - 197V
EletroHarmonix 12AX7 - 216V

It's a good idea to check the V1 voltages when changing tubes, otherwise you may blame a bad tone on the tube when in reality the operating point has changed and the tube is running in one of the 'bad zones'.

Hope this info is useful.

Alan.

[Normster]

Interesting observation, Alan. Now you've got me wondering if the JJ EC83s (I thought were harsh) were running too hot. I guess to really blueprint the circuit, you would also have to measure the bias and adjust the cathode resistor as well?

['67_Plexi]

Yeah from a purely engineering standpoint you can certainly get a read on how well a tube's operating in the circuit by measuring the plate voltage and cathode voltage, providing you have a reference point from a known 'in spec' tube. Also bear in mind the upstream current demands from the other tubes affect the operating voltage of the downstream tubes.
I've seen situations where people have installed an efx loop in an amp and haven't given consideration to the effect of the current demand
from the additional tube. If this extra tube is a 12AT7 or a 12AU7 then the current demands are even higher. Throwing in a 12AU7 efx loop buffer tube can lower the plate voltages on the V1 tube by up to 80V if the dropping resistors aren't adjusted.

On the subject of tubes, I buy 12AX7's in quantities of 50 to 100 depending on how many jobs I have on at the time. I see trends all the time. A few months ago JJ's were coming through pretty consistent, but the last batch I had a lot more 'rogues' appeared. My recent order were Tung Sol reproductions and these were OK on test, but I had an alarming amount with abnormal noise issues when used in critcal gain stages. You just can't win with modern preamp tubes. EH are just plain awful at the moment. On two occasions I've even bought tested and matched tubes right off the bat to try to eliminate my own testing. It didn't work. I've now defaulted back to buying and testing myself. I usually end up discarding between 10 and 15%. That sucks, but in fairness it has been steadily improving over the past 5 years.

[dogears]

That is what over two years of daily dedication, soldering, testing, A/Bing, retesting, up all night, and pissing your wife off help you to attain...

Dogears....is totally correct

[dogears]

Hey Alan,

You have a nice tester? I have a whole bunch of old pulls..... Can you help me test them???

Thanks!!!

[Darkbluemurder]

Very interesting - I guess this could be the key why my modded Concert (70s style) sounds harsh. I haven't measured voltages yet. The schematic shows a B+ at V1 of 300 V which however should be correct by looking at the various schematics. However, the B+ at V2 is shown at 350 V whereas it should be close to 300 V. I guess I need to swap some resistors in the power supply chain.

This is why I love this site. I could probably have gone on forever tweaking in the wrong places ...

['67_Plexi]

Dogears, by all means. Email me at fenflame@comcast.net.

I have an Avo CT160. http://www.duncanamps.com/technical/ct160.html
I'm originally from England and came over to the US in '02. I got this tester from where I worked when we finally stopped servicing tube powered lab equipment in the 80's. It originally came from the British Army and was made in the early 60's.
I calibrated it last year and it was still within tolerance and didn't need any adjustments. It's a great tester.

[seulb-rd]

The point I'm getting to is the fact that the preamp tubes you use play a big factor in this. The 'book' test current draw for a 12AX7 is 1.2mA per triode.

very interesting. where does the "book test current" data come from? can you use an anode current vs voltage graph and extrapolate from there? can knowing this for a tube help find the sweet spot for that particular type of tube. would this work for (preamp) pentodes as well?

['67_Plexi]

I was actually talking about 12AX7 preamp tubes.

Take a look at this link http://tdsl.duncanamps.com/show.php?des=12AX7

Go down to 'Application Data' and look under 'Anode Voltage=250'
You can see it states Ia=1.2mA, this is the 'text book' anode current for 250V. This also corresponds with the test settings used on the AVO CT160 tube tester to test a 12AX7.

[tonelab2]

I have read similar on different sites over the years. The common view is todays 12ax7 isn't as good as the old ones.Apparently hard to find a make of tube that consistantly comes in at 1.2mA or close. From what I understand the mA rating sets the gain of the tube, lower mA lower gain ect. While the transconductance rating reflects the recovery rate of the tube, higher MC slower recovery( after a note is picked). Is this correct or am I talking out of my butt?
Recently sold an Avo 160 (ex navy). Wish I hadn't . I have other MC and Em testers such as a TV2, Hickok 536, Precise 111, RCA 110a, Jackson 658t, bla,bla but the Avo is the ODS of the tester world.

[BobW]

67_Plexi, thanks for your research. Based on your observations using the Mullard standard at 201 VDC and 1.2 mA, it would be interesting to further determine what Vp and Ip works best based on the manufacturer. I know that's a lot more research, but even with a 10 to 15% rejection, it may provide a window to adjust plate resistor values using trimmer pots located at each plate. I assume your Ip @ 1.2mA is based on a no input signal? Thanks again for the research. Bob

[llemtt]

Current production 12AX7s usually have a lower transconductance compared to the original "reference" tubes, this means that with same rp/rk values the idle current will be lower and plate voltage will be higher (imho this also sounds worst).

To me this also explains why HAD increased plate values along the years trying to mantain the best sounding idle currents and plate voltages, as Dogears perfecly says there's a sweet spot in plate voltage where the sound is neither dull nor harsh, go a little over/under and a fantastic amp becomes like many others.

Considering the goop on the preamp side, the very subtle variation in sound between different rp/rk values, I suggest the PS drop resistors (and snubbers which aren't under goo...) to be the HAD most productive tweaking area, at least during normal retubing maintenances.

Given that mu is comparable between old and new production tubes the internal plate resistance turns out to be higher in new tubes but I can't say if this is good or bad.

teo

[Smitty]

We can use ohms law to determine idle current. E=IR or E/R=I

Let's take 1997 HODS R1.2 because it's handy. The input stage idle current can be determined two ways. The plate resistor voltage drop (Ebb - Eb) divided by the plate resistor. (323 - 209.6)/220K = 0.52 ma. Or the drop across Rk divided by its value. Again, 0.52 ma. For the four gain stages the numbers are

V1A Rp 0.52 ma Rk 0.52 ma

V1B Rp 0.78 ma Rk 0.54 ma (discrepancy)

V2A Rp 0.74 ma Rk 0.75 ma

V2B Rp 1.10 ma Rk 1.13 ma

It looks like some gain stages are biased more toward cutoff than clipping/saturation. The one gain stage that is biased more toward clipping (actually half way between cutoff and clipping) has a large current limiting resistor that acts like a limiter when to tube goes in grid conductance. It also darkens the sound a little because of miller capacitance.

If you can look at this stuff on a scope (a cheap one does the trick) it starts to make sense. The idea is to avoid sharp edges on the clipping side.

[BobW]

Calculating IP based on Vp, and/or using a scope or spectrum analyzer can help get in the tone ballpark, but as many here at the Amp Garage agree, it's always best to trust your ears. Obviously, calculating the Rp and Rk values alone assumes all tubes from a particular manufacturer are the same. Clearly they are not. It was simply suggested to determine (by calculating and listening) the Rp and/or Rk variance based on a per manufacturer basis. Knowing this delta then enables additional potentiometers, permanently located at Rp and/or Rk, to fine tune the stage(s) by trusting your ears. Dogears that is.

[llemtt]

Hi Smitty

V1B rk is obviously wrong in 1997 HODS R1.2, in fact it's 2.2k, this explains discrepancy.

Regarding cutoff/saturation you are not completely correct, take a bunch of plate curves and draw load lines just to see that. Using the standard rp/rk ratio of 66.666... the triode is always biased almost in the middle between cutoff and saturation following the empirical 1/3 1/3 1/3 rule (or 1/3 2/3 rule if you like) as suggested from the theoretical grounds of the triode transfer function. There is plenty of basic tube books explaining this.

Consider also that all the stages are cathode bypassed so in a dynamic regime, especially reaching cutoff/saturation, the operating point is shifted significantly and it's difficult to predict with the usual small signal modelling.

An example of a really skewed operating point is the typical pre "cathode-follower" stage in the JCM800/Soldano/Boogie DR/... kind of amplifier, with 100k/10k in JCM and even up to 33k on the rk in Soldanos and others, and without bypass, the stage is really skewed toward cutoff thus contributing a lot of second harmonic distortion into the following stage: but that is another sound!!

Then I completely agree with Bob regading tubes: each brand is different and also within a brand you can experience different tubes. I am planning to build a computerized curve tracer to manage tube that sounds ok and tubes that don't

my .02
teo