• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

How to determine current requirements for output tube from the driver

My post 31: At times you don't realize what the technical details mean.
Your tube Aa has a plate impedance of 30K Ohm. For a linear response you need to load it with (simplified) 30K x 3 = 90K Ohm. Your Monolith IT is ca. 25K Ohm impedance. (simplified: 40H = 5K Ohm) The 25K Ohm IT gives the audio bass range near no load, so you get no linear frequence response. With your 450µF bypass-cap on GM70 1K cathode resistor you have build a bad bass enhancer- you hear it. At the input tube you cut the bass freq. off and at the output tube you try the opposite. A linear bypass-cap value would be ca. 100µF.
A 90K Ohm IT (ca. 720H with airgap) could not be done for linear frequence response. You should have asked 50AE . Ca. 8K Ohm plate impedance (triodes) is the limit where you can use inductive loads. (plate chokes and IT).
Why have you not asked the forum here ?
A EML 30A loaded with your Monolith IT would work good and linear at 8mA. But these modern chzech tubes sound different than a Aa from the 1930ies.
Ask Ale / Moglia, he has used both, the Aa and the EML 20A. The 30A should sound similar to the 20A.
 
The 20A used to be available as both solid plate and mesh, apparently now only available as a mesh plate according to the website. I have had 20AM (mesh) and 20A (solid plate) in the past. I had a lot of difficulty with the 20AM getting noisy over time in some line level applications, and EML specifically now recommends that it not be used in those applications. IMO the 20B is a suitable and more rugged alternative which has significantly lower RP, higher dissipation ratings and from my direct experience is very reliable. (It is also extremely quiet over a long life time so good for line level stuff where I use them.)

I agree with your assessment on the 30A 8mA plate current drive for GM70, it is just adequate. With a 20B you can run 25mA - 30mA (or more!) at 400V and have an extremely linear driver that would also allow limited excursions into class A2 if more power is needed. (If you do this winding DCR needs to be low)
 
Last edited:
  • Like
Reactions: Hifihunter
I do have a 440V secondary on my driver power transformer. Running the numbers on PSUD with a RGN2004 rectifier + a 0.22uF 1st cap in my cLCLC, this produces 400V B+. So, we are good to go from the B+ perspective. 400V @ -10V gives 25-27mA. 400V @ -8V gives 32-32V.

Kevin-any thoughts on how the 20B sounded in your system? Any insights into sonics would be appreciated.
 
  • Like
Reactions: kevinkr
The 20 series in general are the most linear DHTs I have ever encountered. I highly recommend the 20B if you are looking for detailed, accurate sound - in a good implementation they are extremely transparent and add or subtract almost nothing from the sound. I use them at 25mA and 390V (fixed bias) in a line stage application, the current tubes have about 2500 hours of use and have not drifted significantlyy in parameters over that time.

I would also recommend considering the EML520B as your power tube, it will make for a much more efficient and likely reliable amplifier - power output in A1 will be comparable. (Around 15W is feasible in A1)

After 11 years I am considering a new amplifier design with EML520B for my system bass amplifiers, this will reduce power consumption by almost 50% and provide roughly the same output power. I also will not have to change the tubes every 12 months as I now do.
 
Although the 8mA limit of the IT is somewhat restrictive, the 25k anode load it presents is very desirable for the 20B or 30A.

One way to retain it, and still fix the drive to the output tube: Add a MOS source follower. This can be designed to reduce the driving impedance for the GM 70 grid to a few Ohms , but present the IT with a load of >300k plus a few pF.

They are pretty transparent to my mind; but in any case the improvement in the grip on the GM-70 grid will most likely more than make up for any loss. You pick up a little guiltless A2 while you are there, which might help in the peaks.

It's a relatively low cost update, certainly much lower than a premium IT.
 
I would not recommend running the 20B at 8mA, options would include using the 8mA IT in a parafeed circuit with a cascode CCS on the plate of the 20B, but supply voltage would need to be much higher. (700 - 800V)

I think in a case like this the transformer (8mA/25K) is targeted at something like the 6J5. This would work, and you could add powerdrive (mosfet source follower) if you want class A2, otherwise this should work relatively well. You should also talk to Yves about what other tubes he thinks would work well with this IT.
 
Thanks Kevin & Rod. Very helpful. At the time I was working with Yves on the IT, we targeted the REN904. I am going to give this a try, as it should produce nearly 10mA. Not perfect, but certainly an improvement. If this appears to being going in the right direction, then I can splurge on a pair of 20B or 30A.
 
  • Like
Reactions: kevinkr
Why don't people use a cap at the amp input these daze?
Most of these circuits are AC coupled, what benefit does a DC coupled input confer?
Should be a cap at the input connexion to protect the input tube, especially if it is fixed bias. 😀
 
If the input tube is fixed bias by definition you have to have an input capacitor or you may effectively short the bias to ground via the connected component. I don't otherwise use them and after 50 years have yet to see or experience an input tube that failed due to DC on the input. (They mostly just wear out due to excessive use.. lol)

These days I have input transformers and fixed bias on all of my amps except the one in my home office.
 
  • Like
Reactions: Technical
THX for your response Kevin, but it does not answer the question, why do people omit the input cap in their power amp.
I thought we would get a few opinions & reasons why. Perhaps volk are simply doing what they see others have done.

During the so-called Golden Age of vacuum tubes the only amp most people would ever see without
an input cap was to an Op-Amp. Or a low level phono preamp 🙂
 
  • Like
Reactions: kevinkr
THX for your response Kevin, but it does not answer the question, why do people omit the input cap in their power amp.
I thought we would get a few opinions & reasons why. Perhaps volk are simply doing what they see others have done.

During the so-called Golden Age of vacuum tubes the only amp most people would ever see without
an input cap was to an Op-Amp. Or a low level phono preamp 🙂
The Citation II, QuadII, Dynaco ST-70, MKII, MKIII and many other amplifiers did not use the input coupling capacitor. This was likely as much cost related as anything.

The Marantz 8B did use input capacitors as part of a 2nd order HPF, but also provided a provision to bypass it by plugging directly into the "lab" input and many people did.

The cap is there primarily to block DC, or to limit LF response in the amplifier, if there was no perceived technical need or benefit the part would likely have been left out. Tubes were relatively cheap and unless there was a significant risk of damage from the low DC levels likely to be present on the output of most hifi gear the part could be justifiably left out.

I wasn't around designing stuff in that time frame - I am old, but not that old. LOL Suffice it to say that not everyone felt they were necessary, and engineering design is always a matter of choice and compromise, many of the choices are reflecting preferences of the designer themselves based on experience, opinion and design philosophy. Worst case the bean counter comes to you and asks (or tells) you to cut cost wherever possible.
 
  • Like
Reactions: Hifihunter
Attached two old snapshots from the spectrumanalyzer. The EMl 30A is loaded by ISO NC 20F interstage transformer, 1 kHz signal secondaries open during measurements. The first picture shows the spectrum at 116 V rms output, The second shows the spectrum at 77,3 V rms output. Tube biased at 400v@ 20ma.
I used the EMl 30a/ISO Nc20F comboas driver for the GM-70 some years ago, but last year I changed the drivertube to the EML12B (Gain 14.5 When transformer loaded). Changed partly because the 12B has lower internal resistance than the 30a Which is a bit on the high side for the ISO interstage transformer. The Iso interstage was used because I already had them. I plan to replace the iso with a pair of nano-x later this year. Suggest you ask Yves/Ward about which "driver current window" your interstages are comfy with. As you up the current in the EML 30a internal resistance go down. I have no experience with the EMl30a running at such low current as 8ma. At the working point I used it, I think it is excellent!
 

Attachments

  • eml30aisonc20f21khz116vrms.jpg
    eml30aisonc20f21khz116vrms.jpg
    69.3 KB · Views: 95
  • eml30aisonc20f21khz77,3vrms.jpg
    eml30aisonc20f21khz77,3vrms.jpg
    68 KB · Views: 98
  • Like
Reactions: Rod Coleman
Pat!
I'm late in the party ... but ... if you use our Monolith SX-11 5k OPT in GM70 amp, firstly ask Monolith!
It's "only" 400V 10W transformer, so I'm not sure that using it with 1.2kV peek anode voltage is safe.

BTW with these parameters (in your first post) the output would be over 13W, so core may saturate even at 40-50Hz!