• Increase font size
  • Default font size
  • Decrease font size

ward's 150W SE amp blog part 3

AddThis Social Bookmark Button

Moving towards more lethal conditions

As suggested in the previous article, I wanted to apply substantially higher voltages at the plate of the 833A / TB3-1000 tube. There are a lot of things to be gained such as a better damping factor, but also a driver stage that does not need to rely on a high current solid state device. Some quick calculations proved that a 2KV setting would deliver a good compromise. Yves supplied me with the appropriate power transformers and I got things moving. I remember an article from John Camille describing a shunt regulated 211 single ended amplifier (Kyrie) that stated that operating the 211 at high plate voltage (1250V) proved to be essential to deliver the goods in subjective terms . I put 2KV on the plate without changing anything else, and indeed, I have to admit that the sound clarfied, there was a clear sparkle in the sound that was not there on the low plate voltage setting. Next to that, I increased the primary impedance to a higher level by replacing the original 5K bobbin with a 12 K version.  Running at around 140mA delivered the maximum power of around 150W.


Driver schematics for the output tube: using a tube as a follower

Looking at the grid characteristics learned me that I would need to supply around 60mAp to the grid of the output tube in order to enable the full potential of the tube. This means that there were plenty of suitable candidates. At first, I tested triodes and/or triode wired pentodes or beam tertrodes. Replacing the mosfet with a tube immediately improved the sonics of the amplifier, the sound being more agile and more microdetail was present. Perhaps it has to do with the rather high input capacitance of the solid state device I used. It is well known that loading a tube stage with a high capacitance device is not the best option soundwise. I made some sort of universal connector to swap driver tubes. I started with one of my favorites,  black plate NOS 6AS7G  which rewarded me with a very nice subjective performance. However, these kind of tubes are kwown to be prone to arc over and since the driver tube was dc coupled to the grid of the output tube ... It was time to source a good replacement. I evaluated many tubes such as 6C33C, EL34, KT88 (NOS Gold Lion), EL156 (octal version), 8417 and even a vintage WE 300B was tested. I settled on a new production triode strapped EL34 running at 80 mA. 8417 was very very good, however quite difficulot to source. Another benefit of the high operating current condition is that you obtain the full transconductance of the tube being 12mA/V. Being  configured as a follower, this means that a low 80R output resistance can be accomplished and the power tube is driven in a stiff manner.

Looks even better this way :-)

Different kind of tubes I evaluated

EL34 with zirconium getter


Pentode connected driver stage

I have built quite a lot of mu stages in the past. Allan Kimmel (inventor of the mu-stage) goes into great lengths promoting a pentode as a follower. The reason being the fact that there exist quite a lot of high transconductance pentodes and that the transconductance is independant of the voltage applied to the tube if operated in it's intended region. Another benefit is the much lower plate to grid parasitic capacitance. Installing a few switches did the job. Well, it clearly sounded different. There was more muscle in the sound mix. However, the triode configuration proved better in the ambient retrieval of the recording. To me, it all depends on personal tatste and can be used to voice the amplifier and it's for free: the flip of a switch does the job.


driver PCB I developed to enable A2 operation

13E1 high gm tube 

stuffed PCB board equipped with connector


The Input stage and the technical performance of the driver stage

Because of the much reduced capacitance presented by the tube based follower stage, I was able to pick tubes with a higher amplification factor such as some exotic NOS 6SL7 and ECC35. I have been using this tube for quite a long time in my amplifier with great succes. The output resistance of the SRPP configuration proved to be 25K as it should be. The complete driver stage should have a  bandwidth of over 300 kHz because of the low input capacitance of the pentode strapped EL34 follower. I verified the bandwidth of the complete driver stage and to my surpise I got no better than 100 kHz. Not bad at all but substantially less than my expectations, ..calculations. Swapping the 6SL7 with a Red Base 5692 resulted in 350 kHz and I am very satisfied with the sound and technical performance. I still  want to resolve the mystery of the extra 35pF that I am not able to pinpoint. An alternative would be the application of a cascode based current source using the exteremely low capacitance LND150 mosfet as developed by Gary Pimm. This could give me extra amplifcation while still delivering frequency bandwdth. I plan to investigate this in the future.

Red Base 5692

CCS 8A in progress 

Mullard ECC35


The full schematic phase 2

The schematic reveals a fairly simple topology.  

the schematic of the phase 2 833A - TB3/1000 SE amplifier

Time to order thick anodised front plates?

At the end of my previous article, I had some questions that I wanted to resolve. Many of them have been answered. In the meanwhile, I have been rewarded with a nice sounding amplifier. Is there still room for another iteration, or should I clean up the amplifier? moving it from this highly experimental state towards a more commercial look boosting thick shiny front plates. Well, I developed an 8A constant current source PCB to power the mighty filaments. I also have a bunch of 3B28 Xenon  recifiers that I would like to use and I would like to play a little bit with different transformer materials. Currently, I am using  192 kHz usb audio as reference material!  As you see , I am not yet ready ...stay tuned,

to the first page