I have a question applying to design of 2-stage low-power DHT amplifiers:
Getting enough gain from only two stages to me means one of two things: 1) using a high-mu driver tube (like Telefunken KC3), or 2) stepping up the signal by means of a transformer.
Now, high-mu DHTs, KC3 in particular, are hard to come by and you must be ready to bring $$$ if you find them, while "typical pre-amp tubes" like the 26 or 12A with mu around 8 but higher current handling, come relatively cheaply and can be stepped up in an 1:2 interstage for instance, and thus provide decent gain while still being able to deliver enough current for a suitable output tube, the 71A is on my mind.
Does anyone have experience on the matter and wish to share their POV? pros and cons of using high-mu tubes vs. transformer stepup?
Thanks
EDIT: I should mention that with high-mu DHTs one also must deal with high Rp and thus choke/transformer loading is out of the question as inductance requirements raises ridiculusly high, so RC-loading would be a con for the high-mu tubes I think..
Getting enough gain from only two stages to me means one of two things: 1) using a high-mu driver tube (like Telefunken KC3), or 2) stepping up the signal by means of a transformer.
Now, high-mu DHTs, KC3 in particular, are hard to come by and you must be ready to bring $$$ if you find them, while "typical pre-amp tubes" like the 26 or 12A with mu around 8 but higher current handling, come relatively cheaply and can be stepped up in an 1:2 interstage for instance, and thus provide decent gain while still being able to deliver enough current for a suitable output tube, the 71A is on my mind.
Does anyone have experience on the matter and wish to share their POV? pros and cons of using high-mu tubes vs. transformer stepup?
Thanks
EDIT: I should mention that with high-mu DHTs one also must deal with high Rp and thus choke/transformer loading is out of the question as inductance requirements raises ridiculusly high, so RC-loading would be a con for the high-mu tubes I think..
I have used both approaches driving 2A3 and 45 type output tubes.
I have so far only used 1:1 transformers because of my concern for source impedance.
Say for example you used a 27 which has a plate resistance of about 8K, a 1:2 stepup transformer would give you a 32K source impedance (remember the impedance transformation ratio is the square of the turns ratio or voltage transformation ratio) which is not going to result in very good high frequency response to put it mildly when you consider the miller capacitance of the output tube and likely strays on secondary side of that transformer.
Consider a 5842 or 5687 driving a 1:1 transformer to your output tube, source impedance will be low and both have moderately high mu so sensitivity will be reasonable in both cases.
Hey I love the 26, 27, 30, 56 and others, but their low mu and relatively high rp can be a design liability in driver stages and asking them to swing large voltages on their grids in order to get the required plate swing can result in poor linearity as well. (Which might be an asset if the resulting spectra matches the output tube's well aka distortion cancellation.)
Just my thoughts..
Kevin
😀
I have so far only used 1:1 transformers because of my concern for source impedance.
Say for example you used a 27 which has a plate resistance of about 8K, a 1:2 stepup transformer would give you a 32K source impedance (remember the impedance transformation ratio is the square of the turns ratio or voltage transformation ratio) which is not going to result in very good high frequency response to put it mildly when you consider the miller capacitance of the output tube and likely strays on secondary side of that transformer.
Consider a 5842 or 5687 driving a 1:1 transformer to your output tube, source impedance will be low and both have moderately high mu so sensitivity will be reasonable in both cases.
Hey I love the 26, 27, 30, 56 and others, but their low mu and relatively high rp can be a design liability in driver stages and asking them to swing large voltages on their grids in order to get the required plate swing can result in poor linearity as well. (Which might be an asset if the resulting spectra matches the output tube's well aka distortion cancellation.)
Just my thoughts..
Kevin
😀
Another thought, the 26 can be very microphonic as well. I use them in my pre-amplifier design where the 5:1 stepdown transformer reduces the source impedance to roughly a few hundred ohms, and gain is very limited as well. If you are curious see my website at www.kta-hifi.net, go to pre-amplifiers and click on the topmost link (26dht). The design was published not long ago in the positive feedback online magazine.
This will give you some idea of the problems involved with transformer coupled dht circuitry.
Kevin
This will give you some idea of the problems involved with transformer coupled dht circuitry.
Kevin
Finding a high Mu DHT is not easy. Finding a high Mu DHT that is not microphonic is harder.
Kevin mentioned the 5842. It is not a DHT, but I have used it successfully in several 2 stage amplifiers including 45, 2A3, 300B, and several triode wired pentodes. The key to getting good gain is to use a very high impedance load. I use a CCS and a mosfet buffer to isolate the driver from its load.
I know that many will want to build a pure tube circuit, which is possible, but this works very good. If you are interested, the schematics are on my web site.
http://www.tubelab.com/TubelabSE.htm
http://www.tubelab.com/powerdrive.htm
Kevin mentioned the 5842. It is not a DHT, but I have used it successfully in several 2 stage amplifiers including 45, 2A3, 300B, and several triode wired pentodes. The key to getting good gain is to use a very high impedance load. I use a CCS and a mosfet buffer to isolate the driver from its load.
I know that many will want to build a pure tube circuit, which is possible, but this works very good. If you are interested, the schematics are on my web site.
http://www.tubelab.com/TubelabSE.htm
http://www.tubelab.com/powerdrive.htm
Thanks for pointing out the source impedance issue, Kevin. It completely slipped my attention. Using IDHT drivers are not of interrest at the moment, as I've been playing with high-mu IDHTs in spuds for quite a while. I can drive my speakers with 6C45Pi or similar and still have headroom like a fart in free air so to say, so if I am to build myself a DHT amplifier it only makes sense to me if I can get rid of the IDHT, not just adding a DHT stage for the sake of it.
I've gone over the numbers again and again, and in the end I don't think I'll need that much gain from the driver tube afterall, so IT-loaded (1:1) '26 (mu=8.3) into a 5K-loaded 71A (mu=3) output tube would work. I need some time to decide on the '26 though. I'm not sure it is my kind of tube, and I know I am not going to power those 1.5V/1.05A filaments with batteries. 😱
I have heard rumors about the 26 and the difficulties of implementing it due to it's notorious microphonics and strange heater requirement.. All senses is telling me I really shouldn't be messing with that tube, but I think I'll give it a shot regardless, knowing that desperate attempts to battle the hum will probably cause a bad headache
If all attempts to quiet the hum fails, I can always substitute the 12A if I keep a spare heater winding for it.
I've gone over the numbers again and again, and in the end I don't think I'll need that much gain from the driver tube afterall, so IT-loaded (1:1) '26 (mu=8.3) into a 5K-loaded 71A (mu=3) output tube would work. I need some time to decide on the '26 though. I'm not sure it is my kind of tube, and I know I am not going to power those 1.5V/1.05A filaments with batteries. 😱
I have heard rumors about the 26 and the difficulties of implementing it due to it's notorious microphonics and strange heater requirement.. All senses is telling me I really shouldn't be messing with that tube, but I think I'll give it a shot regardless, knowing that desperate attempts to battle the hum will probably cause a bad headache
If all attempts to quiet the hum fails, I can always substitute the 12A if I keep a spare heater winding for it.
Kevin,
I am curious about your experience with the '26. You use rectified filaments wich probably introduces hum to the '26, but you are also stepping down that hum on the output with the rest of the signal..
I would want to IT couple 1:1 the '26 to a 71A output tube, and would like rectified filaments on both tubes. My worry is hum and microphonics. Regarding microphonics, I'll be using teflon sockets to provide mechanical damping, and with a low amplification factor of only 3x following the '26, would you think microphonics will be a problem provided I do some tweaking to get it right?
Hum is the second and probably largest issue, but not for me I think. The amp will do duty from 350Hz upwards, so I believe I can tolerate quite some hum as I doubt it will be audible on my horns. What do you think about that issue? Here, the low amplification factor following the '26 also comes into play.. the hum will only be amplified 3x before being stepped down in a 5K:16R OPT.
EDIT: typos
I am curious about your experience with the '26. You use rectified filaments wich probably introduces hum to the '26, but you are also stepping down that hum on the output with the rest of the signal..
I would want to IT couple 1:1 the '26 to a 71A output tube, and would like rectified filaments on both tubes. My worry is hum and microphonics. Regarding microphonics, I'll be using teflon sockets to provide mechanical damping, and with a low amplification factor of only 3x following the '26, would you think microphonics will be a problem provided I do some tweaking to get it right?
Hum is the second and probably largest issue, but not for me I think. The amp will do duty from 350Hz upwards, so I believe I can tolerate quite some hum as I doubt it will be audible on my horns. What do you think about that issue? Here, the low amplification factor following the '26 also comes into play.. the hum will only be amplified 3x before being stepped down in a 5K:16R OPT.
EDIT: typos
Hi Rocky,
Actually with decently filtered/regulated dc filament supplies hum is not going to much of a problem at all from this source - the problems I had related entirely to coupling into the output transformer, and this will be possibly be an issue with your driver interstage transformer if the power transformer(s) are in close proximity.
Incidentally the 26 is a great sounding tube and well worth the effort to get running properly.
The overall broadband noise coming out of my transformer coupled pre-amp is < 100uVrms over a 22kHz measurement bandwidth, and there is no measurable or observable hum.
I expect that sensitivity will be quite low in this configuration, as a rough guess sensitivity for full output might be around 3.75Vrms, which may also exceed the linear swing possible at the grid of the 26. (too close to cut off at one end, grid current at other assuming -9V grid bias ?)
An interstage with a ratio of 1:1.5 might be a reasonable compromise, raising the source impedance to 20K, then again...
I still recommend the 5842/417A with 1:1 interstage tranny - you will get great linearity, very low source Z and lots of current capability.
Kevin
Actually with decently filtered/regulated dc filament supplies hum is not going to much of a problem at all from this source - the problems I had related entirely to coupling into the output transformer, and this will be possibly be an issue with your driver interstage transformer if the power transformer(s) are in close proximity.
Incidentally the 26 is a great sounding tube and well worth the effort to get running properly.
The overall broadband noise coming out of my transformer coupled pre-amp is < 100uVrms over a 22kHz measurement bandwidth, and there is no measurable or observable hum.
I expect that sensitivity will be quite low in this configuration, as a rough guess sensitivity for full output might be around 3.75Vrms, which may also exceed the linear swing possible at the grid of the 26. (too close to cut off at one end, grid current at other assuming -9V grid bias ?)
An interstage with a ratio of 1:1.5 might be a reasonable compromise, raising the source impedance to 20K, then again...
I still recommend the 5842/417A with 1:1 interstage tranny - you will get great linearity, very low source Z and lots of current capability.
Kevin
There are a couple of other solutions as well. You could use a pentode stage to drive the DHT. This is the Western Electric solution. Suitable pentodes would include: 12BY7A, 6AG7, or the 6BQ5. All have enough "moxie" to drive the DHT. You could also include a common drain MOSFET driver stage. That, too, can certainly drive a DHT with ease.
Kevin,
PSU/mains will not be anywhere nearby. I've just made a nice solidworks layout for dual mono with external PSU, and I think that'll be the way to go. You are right about the sensitivity, it will be low, but then again my power requirements will not lay anywhere near the 600mW-700mW this beast can put out at 4Vrms swing on the grid of the 26. I'll most likely never pull more than 100mW out of it, and that is in extreme cases only, the speaker sensitivity is above 114dB/W/m, so even with my 2.5Vrms output DAC, I think passive preamplification will be an option no matter how crazy it sounds..
I think I'll make a deal on AE's 50H 1:1 ITs at 98eur/pair (special offer). 50Hy is low for the 26, but given required bandwidth it'll work great for my horns. My existing amp is likely to get scavenged for input and outrput iron, Lundahl and Tango.
Sorry guys, both IDHTs and pentodes are probably great drivers, but they are not an element in this experiment. I do not wish to use them even if they perform great, I'm just stubborn on this matter. This will be a DHT experiment or no experiment at all.
PSU/mains will not be anywhere nearby. I've just made a nice solidworks layout for dual mono with external PSU, and I think that'll be the way to go. You are right about the sensitivity, it will be low, but then again my power requirements will not lay anywhere near the 600mW-700mW this beast can put out at 4Vrms swing on the grid of the 26. I'll most likely never pull more than 100mW out of it, and that is in extreme cases only, the speaker sensitivity is above 114dB/W/m, so even with my 2.5Vrms output DAC, I think passive preamplification will be an option no matter how crazy it sounds..
I think I'll make a deal on AE's 50H 1:1 ITs at 98eur/pair (special offer). 50Hy is low for the 26, but given required bandwidth it'll work great for my horns. My existing amp is likely to get scavenged for input and outrput iron, Lundahl and Tango.
Sorry guys, both IDHTs and pentodes are probably great drivers, but they are not an element in this experiment. I do not wish to use them even if they perform great, I'm just stubborn on this matter. This will be a DHT experiment or no experiment at all.
Everyone is entitled to their own opinions. Just let us know how it works out so that we can all learn something. Very few of us have speakers with this sensitivity level, so you are in a unique situation.
Konnichiwa,
It might help to read Ciro Marzio's article on the Katelelo Amplifier, a very similar design. It may further be wise to read VERY CAREFULLY what he writes on IT Coupling (I concur BTW).
http://www.audiodesignguide.com/se/katelelo.html
FWIW, my current project is
WE 102L resistor loaded ->
WE101D/WE205D/300B/10Y et al/71A/45/2A3 (configurable according to desire, 12K Resistor Load with 400/500V +B) ->
Output Valve which can be 300B or "super 300B" for > 15W Pout
A lot of stuff similar to the Katelelo, but as the main difference a simple single heater supply (DC) and solid state HT to fit the chassis I have.
Sayonara
Rocky said:
It might help to read Ciro Marzio's article on the Katelelo Amplifier, a very similar design. It may further be wise to read VERY CAREFULLY what he writes on IT Coupling (I concur BTW).
http://www.audiodesignguide.com/se/katelelo.html
FWIW, my current project is
WE 102L resistor loaded ->
WE101D/WE205D/300B/10Y et al/71A/45/2A3 (configurable according to desire, 12K Resistor Load with 400/500V +B) ->
Output Valve which can be 300B or "super 300B" for > 15W Pout
A lot of stuff similar to the Katelelo, but as the main difference a simple single heater supply (DC) and solid state HT to fit the chassis I have.
Sayonara
Kuei,
You are referring to his points of view on expensive interstages vs cheaper inductive anode load with coupling capacitor? I have less experience with interstages, so I ask for advice on some possibilities. My options as I see it, are amongst the following three choices:
1) For 98 EUR I can get a pair of potted AE interstages, 1:1, double C-core with 30H primary inductance. LF response will have premature rolloff, but I can hardly see use for this amp powering anything except my 350Hz horns, where the 30H load will be just fine.
2) Spending a bit more, I can get unpotted Lundahl LL1660/10mA double C-core interstage wich I can wire for 1:1.125 with 130H of inductance.
3) For just above double the cost of the AE transformers, I can get LL1668 200H/25mA plate chokes, LL1670 540H/1.2mA grid chokes, and 0.1uF silver mica coupling capacitors. All iron double C-core.
By the way, I lean more and more towards the 12A instead of the 26. I think it will be much easier to work with.
You are referring to his points of view on expensive interstages vs cheaper inductive anode load with coupling capacitor? I have less experience with interstages, so I ask for advice on some possibilities. My options as I see it, are amongst the following three choices:
1) For 98 EUR I can get a pair of potted AE interstages, 1:1, double C-core with 30H primary inductance. LF response will have premature rolloff, but I can hardly see use for this amp powering anything except my 350Hz horns, where the 30H load will be just fine.
2) Spending a bit more, I can get unpotted Lundahl LL1660/10mA double C-core interstage wich I can wire for 1:1.125 with 130H of inductance.
3) For just above double the cost of the AE transformers, I can get LL1668 200H/25mA plate chokes, LL1670 540H/1.2mA grid chokes, and 0.1uF silver mica coupling capacitors. All iron double C-core.
By the way, I lean more and more towards the 12A instead of the 26. I think it will be much easier to work with.
Konnichiwa,
Yes, more or less. I have tried a lot of interstage transformer and I come back to capacitor coupling. The only interstage transformers I found acceptable had a huge nickel core and tolerated no DC, so had to be used parallel feed. Using a gridchoke sounded even better.
Ahh, you did not mention that the amplifier was not full range. In a "midrange only" amplifier IT's are much less of an issue.
Ciao T
Rocky said:
Yes, more or less. I have tried a lot of interstage transformer and I come back to capacitor coupling. The only interstage transformers I found acceptable had a huge nickel core and tolerated no DC, so had to be used parallel feed. Using a gridchoke sounded even better.
Rocky said:
Ahh, you did not mention that the amplifier was not full range. In a "midrange only" amplifier IT's are much less of an issue.
Ciao T
Thanks Kuei.
I'll be gambling for the AE trannies then. I like the fact that they are potted, and they come at a very reasonable cost so I can spend some on teflon sockets, silver wiring, etc without trashing my budget.
One more thing;
I just found out I can source some reasonably cheap SLA batteries, would there be a problem to power both channels (B+) from the same stack of batteries?
Thanks
I'll be gambling for the AE trannies then. I like the fact that they are potted, and they come at a very reasonable cost so I can spend some on teflon sockets, silver wiring, etc without trashing my budget.
One more thing;
I just found out I can source some reasonably cheap SLA batteries, would there be a problem to power both channels (B+) from the same stack of batteries?
Thanks
Looks like this amp took quite another turn than expected, the driver tube is not going to be a DHT at all, sorry to folks who thought this was going to be a pure DHT project, it is not.
I found out I wanted and needed the gain some IDHTs provide, and I am thinking about e810f/7788 (triode wired) as I have them handy. I just got my hands on some step-up TVCs I will use at the DAC input, wich with an 1:4 step up ratio and a CCS-loaded e810f triode with mu of 70 (can this be right?) will provide a gain of about 280 at full volume wich in turns mean I can start looking for other DACs than the TDA1543...
The CCS will be battery biased, I've gotten some 0.2uF/350V/0.3% silver mica capacitors I will use both for coupling and bypass of the lower battery in Pimm's BBMCCCS. I plan to double-load the CCS mu output with an extra parafeed step-down stage for pre out for my bass amps (they have to be tapped here to be controlled by the TVC).
Waiting for some other stuff, a stepped volume switch, some sockets, etc before I can start build. I'll have to get a chassiplate for it too, and that'll first come after christmas holydays. This is how it looks at the moment, comments are welcome.
I found out I wanted and needed the gain some IDHTs provide, and I am thinking about e810f/7788 (triode wired) as I have them handy. I just got my hands on some step-up TVCs I will use at the DAC input, wich with an 1:4 step up ratio and a CCS-loaded e810f triode with mu of 70 (can this be right?) will provide a gain of about 280 at full volume wich in turns mean I can start looking for other DACs than the TDA1543...
The CCS will be battery biased, I've gotten some 0.2uF/350V/0.3% silver mica capacitors I will use both for coupling and bypass of the lower battery in Pimm's BBMCCCS. I plan to double-load the CCS mu output with an extra parafeed step-down stage for pre out for my bass amps (they have to be tapped here to be controlled by the TVC).
Waiting for some other stuff, a stepped volume switch, some sockets, etc before I can start build. I'll have to get a chassiplate for it too, and that'll first come after christmas holydays. This is how it looks at the moment, comments are welcome.
Attachments
I ended up doing this all-DHT in the end.. very similar to the schematic I posted exept the UX-112 serves duty as driver. Brought it to the lab here the other day and did some tuning, it sounded very very good on my grado headphones but there was noticable hum. Everything was powered with lab-supplies though and I really plan for SLA batteries on the filaments, so maybe that'll better things. I am building it into a proper chassis soon, but the IAG chassis I ordered a month ago at iagaudio.com still hasn't arrived.. 🙁 If i'm lucky, and get my IAG chassis soon, I might be able to finish it up within two weeks, before I have to attain a mil excercise upnorth for 3 weeks.. when i get back from that I have four weeks unntil my engineering exams, so not much time for hobbies these days 
Hello Rocky! Hilsen fra Stor Brittania.
funny I should read this today - I was just planning my next DHT amp and laying out componants seeing what space they will take. My source is a DAC with differential out and 1.8v DC on the balanced outputs. This at present goes into a balanced 26 line stage, followed by teflon coupling caps and a simple 12 way stepped attenuator in shunt.
This goes into my amp which is 2C22 in balanced SRPP (same electrically as 6SN7) driving 6AH4 outputs in push pull. Fairly unusual choice of tubes, but very nice sound.
I first used batteries for the 26s which was silent, but then powered them from two bench supplies (Thurlby Thandar) which were nearly silent, and MUCH more convenient. My speakers are 88db, single drivers (Monacor 130AL). In terms of gain, this system is just right for my small/medium room.
My latest thoughts go as follows:
- use two mono amps and forget the line stage
- use CCSs under the diff pairs
- take the balanced signal from the DAC into the 26s, followed by reflon coupling caps and something like a 47k stepped attenuator in shunt. This is what I have and it all works fine.
- following this I could use something with mu = 20. I know the 2C22 stage sounds excellent - almost DHT. My question here is the same as yours - is there a DHT with mu >15. There are some quite obscure possibilities - DCC90, 3B7, 3A5 in doubles, and in singles DC93, DC80, CK5676, 1LE3, 3A/142A, 3A/108 and some others. I can't see myself using any of those, and chums that have used some found them microphonic. So I think here we will stay with 2C22. Other possibilities are 6S4, 76, 6P5GT and such. Maybe ECC40 or E80CC which is about as much as I find myself liking in smaller tubes. And the loctals 7N7, 7AF7, 7A4 etc.
The gain of this three stage setup is OK for me - it's balanced so that helps too - and the Russian teflon coupling caps and stepped shunt attenuator don't interfere too much with the good sound. Keeping everything in one amp simplifies the topology and allows small value coupling caps, since the teflons come in 0.1uF. The DC on the DAC output gets absorbed by the CCS under the 26s.
The output tubes - 6AH4 - have a mu of 8 so I may be in trouble with gain if I use 2a3s which I certainly want to try. I also want to try 71A in the outputs and some others like 1624 and 1626. This is where it gets tricky again - I want to keep the 26s in the first stage - they just sound so good. So if I need more gain I may need to go above mu=20 for the second stage, and that's the question you seem to be asking! Mind, with horns, you gain sensitivity, but then on the other side you are SET. Andy
funny I should read this today - I was just planning my next DHT amp and laying out componants seeing what space they will take. My source is a DAC with differential out and 1.8v DC on the balanced outputs. This at present goes into a balanced 26 line stage, followed by teflon coupling caps and a simple 12 way stepped attenuator in shunt.
This goes into my amp which is 2C22 in balanced SRPP (same electrically as 6SN7) driving 6AH4 outputs in push pull. Fairly unusual choice of tubes, but very nice sound.
I first used batteries for the 26s which was silent, but then powered them from two bench supplies (Thurlby Thandar) which were nearly silent, and MUCH more convenient. My speakers are 88db, single drivers (Monacor 130AL). In terms of gain, this system is just right for my small/medium room.
My latest thoughts go as follows:
- use two mono amps and forget the line stage
- use CCSs under the diff pairs
- take the balanced signal from the DAC into the 26s, followed by reflon coupling caps and something like a 47k stepped attenuator in shunt. This is what I have and it all works fine.
- following this I could use something with mu = 20. I know the 2C22 stage sounds excellent - almost DHT. My question here is the same as yours - is there a DHT with mu >15. There are some quite obscure possibilities - DCC90, 3B7, 3A5 in doubles, and in singles DC93, DC80, CK5676, 1LE3, 3A/142A, 3A/108 and some others. I can't see myself using any of those, and chums that have used some found them microphonic. So I think here we will stay with 2C22. Other possibilities are 6S4, 76, 6P5GT and such. Maybe ECC40 or E80CC which is about as much as I find myself liking in smaller tubes. And the loctals 7N7, 7AF7, 7A4 etc.
The gain of this three stage setup is OK for me - it's balanced so that helps too - and the Russian teflon coupling caps and stepped shunt attenuator don't interfere too much with the good sound. Keeping everything in one amp simplifies the topology and allows small value coupling caps, since the teflons come in 0.1uF. The DC on the DAC output gets absorbed by the CCS under the 26s.
The output tubes - 6AH4 - have a mu of 8 so I may be in trouble with gain if I use 2a3s which I certainly want to try. I also want to try 71A in the outputs and some others like 1624 and 1626. This is where it gets tricky again - I want to keep the 26s in the first stage - they just sound so good. So if I need more gain I may need to go above mu=20 for the second stage, and that's the question you seem to be asking! Mind, with horns, you gain sensitivity, but then on the other side you are SET. Andy
- Status
- Not open for further replies.