This months tube is a very special type with some extraordinary properties, the 6HS5. This is a quite 'modern' tube, developed in the 1960ies. It's application was high voltage regulation for the acceleration voltage of color TV picture tubes. It was intended as a pulse type high voltage regulator.
This tube is a beam triode. This means it has another electrode between grid and plate. Similar to beam power tetrodes, it has a beam forming plate. What makes this tube so unusual are it's extremely high amplification factor of 300 together with a transconductance of 65.000 micromhos! This results in a plate resistance of about 5kOhm which makes the tube usable for LC or transformer coupling. Thus it's full amplification can be utilized.
The 6HS5 has a 12 pin compactron base. The compactron socket was introduced by General Electric. The purpose of the many pins was the possibility to integrate multiple systems within a single envelope. The 6HS5 only uses part of the available 12 pins. Control grid and beam forming plate are brought out to two pins. Only one of them needs to be connected. The beam forming plate can be connected to the cathode at the socket. There is a family of similar tube types which all share almost the same electrical parameters and have the same basing diagram: 6HS5, 6HV5A, 6HZ5, 6JD5. They only differ in the heater current and some have 35W plate dissipation instead of 30. I work mostly with the 6HS5 due to it's rather modest heater current of 1.5A at 6.3V. 30W seems ideal for the usage as an output tube. In single ended confuguration this would yield about 7W. More like 8W with the 6HV5As 35W max plate dissipation. These tubes are available in abundance from almost each reputable tube dealer at ridiculously low prices. Why are they so overlooked by the audio crowd?
Well, nothing comes for free. While the tube has exceptional transconductance paired with a very high mu, it requires a rather high plate voltage to operate. In single ended Class A, over 1000V B+ is needed to achieve sensible results. This is probably the main reason why it is not so commonly used, besides the 'TV tube stigma' which causes many very interesting tubes to be ignored by audio designers.
The curves in the datasheet compare well with my measured results. The picture below shows the plate curves of a actual tube, taken with a curve tracer. This is only a small section of the curves since the tracer only reaches about 350V. As mentioned above plate voltages beyond 1000V are needed for reasonable results. Still this small part of the curves shows that it is very well usable for audio.
Plate voltages of 1200-1300V are still quite reasonable and DIYers who are used to build with transmitting tubes like 211 or 845 can handle this. Another difficulty however is the highish plate resistance. Measurements showed plate resistances above the 5kOhms mentioned above. So an output transformer with high primary impedance would be needed. Very few such transformers exist. Suitable would be the Tango FW20-14S with 14k primary impedance, or the Lundahl LL2735B with 16k primary impedance.
The schematic on the left shows the circuit of a power amp with the 6HS5. Due to the high amplification factor, no driver or input tube is needed. This makes a so called spud amp possible with this tube. The awkward term 'spud amp' has a somewhat funny history. An amp with one tube per channel is a 'one tuber'. Tuber is also an expression for a potato. That's why these amps are also called potato amps sometimes. And spud is just another synonym for potato. So all which is needed is the output transformer, a cathode resistor with appropriate bypass cap and a high voltage supply. The schematic also shows a decoupling choke and the decoupling cap is connected in ultrapath fashion, my favorite topology for transformer coupled stages. The tube operates at about -3 to -3.5V grid bias. So up to 7V peak to peak input voltage is needed for full power out. That's about 2.5V RMS. A bit above the typical line level, so it is preferred to drive such an amp from a pre amp with some gain. Below is a photo of a single ended power amp using one 6HS5 per channel:
This circuit can be easily scaled up to achieve more power by paralleling tubes. A Parallel Single Ended verison with two tubes per channel delivers twice the power:
This scheme can be even scaled further. Below a photo of Quad Parallel Single Ended mono blocks. These deliver solid 30W single ended pure Class A:
Close up photos, showing the glow of the tubes. The blue glow is normal and does not indicate bad tubes:


Thomas
Hi Thomas.
ReplyDeleteThis tube looks like a good candidate to direct drive an electrostatic speaker. Maybe just the treble panel of a hybrid ESL. A differential SRPP pair consisting of four tubes per channel with +/-1200V bipolar power supply might do the job. Filaments will need separate floating power supply. I have been planning to build a direct drive electrostatic headphone amp for Stax. I had 6S4 or 6AH4GT in mind. 6HS5 might be easier to implement with just two stages and enough gain for feedback.
Hi!
DeletePlease keep me updated about your results if you try this
best regards
Thomas
Does exists similar, in small package, weaker?
ReplyDeleteNone I am aware of
DeleteThomas might there be sonically interesting line level applications for this tube at much lower voltages? Maybe 250v based on your curves?
ReplyDeleteno
DeleteHey just wanted to say I think your talented builder and look forward to reading more of your site
ReplyDeleteNow with that said where can I buy these tubes at ridiculous prices? If I can find them for a buck each I would buy 100
Lawrence
If you want ridiculous prices you need to search and be patient. Normally they are offered around $5-$8 which is still ridiculously low for such a piece of technology
DeleteThomas
I would disagree these are not rediculious cheap a dollar a piece is more like it :)
ReplyDeleteI have been tested 6hv5's beam plate connected plate , or connected to cathode . I don't find much of difference with sound. How do you think ,it is better to connect with plate or cathode?
ReplyDeleteThis comment has been removed by the author.
DeleteI never tried beam plate connected to anode.
DeleteOnly connected to cathode
Hi
ReplyDeleteIn article above written by you "The beam forming plate can be connected to the cathode at the socket. ". It is right?
This appears to be the ideal valve to drive a Pair of triode strapped 813.
ReplyDeleteInteresting,the cavity anode design was the last development of the very last units.
The more fragile 8417 was done this way, and also has very high gain.
Hi Thomas,
ReplyDeleteWhat about a PPP amp based on the 6HS5 with a phase splitter trafo at the input? What Lundahl would you recommend for the job? It seems like otherwise we could adapt your SE schematic.
Hi!
DeleteI have not worked with the 6HS5 for a while. It has very high plate impedance so even a PPP amp would require OPTS with very high primary impedance. Maybe the LL1691B would work
BR
Thomas
I am making KT150 CSPP amplifier, using 6HS5 as a driver with anode choke PP drive. CSPP requires quite a high driving voltage so I hope to use 6HS5 600v on plate. I don't know if it is high voltage enough for 6HS5 ? I am going to try.
ReplyDeleteI don't think that will be enough in order to get the output voltage you need
DeleteThis was a very interesting thread.
ReplyDeleteMay be interesting to note it's extremely easy to find a HT source to power a few of these.
Just find a couple of microwave oven transformers.
You can choose to run 230V ones in series to halve the output and/or give a centre tapped +/- supply rails. Very cheap and powerful.
How about a cathod follower to drive speker via more common xfrmr?
ReplyDeleteyou can try that, but you will need a second stage to provide the large drive voltage needed
DeleteI have used the 6HS5 as a dc coupled deflection amplifier, It worked great and we sold a few. The biggest problem was a low capacitance heater winding. Transformer winders would not understand that air in the winding was a must. We got part of the way but never much beyond 5kHz because of the capacitance of the heater winding. We used -2500V to +2500V and the output swung +/- 2000V. the standing current was 11mA, so the tubes dissipated 28W all the time. It was all class A and lots of negative feedback, probably 50dB or more. H J Weedon
ReplyDelete