• 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.

Help on designing full DHT SET power amplifier

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi guys.

i need help on idea and suggestion on building a SET amp

i am thinking of using PTT2S then follow by 45. i am using it to drive a 95db full range speaker. my cdp has around 1v output.

i calculate the gain of PTT2S is around 13.5x
while 45 is round 2.6x

the gain of the output transformer 5k:16ohm is 1:17.5

hence total gain of the two stage with OT is around 13.5*2.6/17.5 = 2x. please correct me if i am wrong

whats the recommended gain for this kind of SET amp? should i chose other tube? any recommendation?

thanks alot
erwin
 
Thanks mr Shalom

nice. thanks for the input. yeah step up transformer is a nice option. could eliminate a stage of tube amplification. superb. any suggestion?

how bout LL1660 or LL1692a using alt Q connection but connect it the other way around? 4.5:1 or 3.5:1 respectively? or proper input transformer is recommended?

thanks,
erwin
 
Last edited:
Hi
Your approach seems a little backward for me. Normally you decide on the output stage operating point. This will give you max output power possible and the signal voltage needed at the output tube grid to drive the tube. I have not worked with the 45 but looking in the RCA data sheet one OP is: anode voltage 275V, idle current 36mA and a grid voltage of -56V. This will give you 2W out.
To drive the 45 to max power(A1) you will need 56Vpk signal at the grid. This is approx 39,5Vrms. If you have 1Vrms from your CDP it means that you need minimum a gain of 39,5 from input to the 45 grid.

/Olof
 
Here's an example of a DHT SET that I use personally. The 26 stage is in a separate preamp but can be in the amp chassis. The signal chassis has to contain the Rod Coleman filament supplies and the heatsinks needed. I have a separate power supply and filament supplies. Note the 26 is in filament bias, and uses a choke input supply with a 18+18v transformer. Andy
 

Attachments

  • 300B O-netics.png
    300B O-netics.png
    61.8 KB · Views: 1,109
hi Olof

thanks alot for the maths. it really help. superb.

hi Andy

Nice. i saw your work and admire it alot. your work is one of the inspiration for me to build my first DHT #26 preamp which sound superb.

Hi Thomas

is there any limitation with input step up transformer approach compare to the 3 stages? or can you tell if there is any sound signature differences in the two approaches?

Hi Bigun

nice link.
Thanks guys for all the inputs

I got few questions:
1. if i am using Ptt2s to drive the final. Ptt2s has internal resistance of 25K. would i still be able to use interstage transformer?
for example Tango nc20f has 5k:5k whats the max Rp for the driver stage? 10k? how bout 25K? or is there any IT suggestion for Rp= 25k?

2. for example tube 26. if we chose the operating point such that the grid voltage is -4.5v. does it mean that the previous stage max rms input is 2.8Vrms?

3. is there any tube suggestion to drive tube 45?

thanks in advance
erwin
 
Last edited:
Hi Erwin,

I prefer less gain stages. If the preamp has low Zout you can take advantage of that by achieving some gain with an input transformer.

I found good input transformers to be very transparent.

As you might have read from my posts, I prefer LCL filament supplies. With that approach a 3 stage DHT amp would get very big and heavy.

Best regards

Thomas
 
Hi Erwin,

I prefer less gain stages. If the preamp has low Zout you can take advantage of that by achieving some gain with an input transformer.

I found good input transformers to be very transparent.

As you might have read from my posts, I prefer LCL filament supplies. With that approach a 3 stage DHT amp would get very big and heavy.

Best regards

Thomas

LCL filament supplies are great, but yes they are bigger and heavier, so I only use choke input on the 26. You run into some serious practical questions with DHT amps - what can you actually put in your signal chassis? Here's the list of what needs to go into a two stage stereo amp with external PSU:
- OPT x 2
- Interstage from driver to finals x 2
- Final tubes (2) plus cathode resistors and polypropylene cathode bypasses
- Driver tubes (2)
- Four DC filament supplies (e.g. Rod Coleman) plus some input caps plus their heatsinks. The heatsinks are serious sizes for 300b, 10Y, 6B4G, 45, 46 etc. You have the option of using AC on the 300b or 2a3 finals or whatever, also the option of using driver tubes with smaller filament current like 112A, 31 or 71A. Those choices eliminate a lot of heatsink, but you have to ask yourself if these give you the sound you want. I'm sticking with 46 after trying a whole bunch of DHTs. You can use LCL but that moves the weight and size into separate filament supplies - filament chokes for tubes like 26 have a significant field around them - move them closer to the 26 and you actually hear the hum going up!
- Some final HT components, like polypropylene caps for driver and final stages and dropper resistor or whatever for the driver stage.

It is physically possible to put the 26 input stage in the amp - it all goes into a chassis 420 x 300mm i.e. standard 2U subrack size (I use subrack parts). But it's a close fit and the signal chassis is then quite heavy. This means that in principle you don't need step down and step up transformers. I actually use a separate 26 preamp with plate chokes and cap out, so 3 stage. This gets the job done in reasonable size chassis.

You have to plan what goes where in some detail with all-DHT amps. Don't buy a chassis you like and have blind faith that what you need will go in it - in most cases it won't. Allow good space for output transformers - Lundahl ones are big and need to go under the chassis unless you use some kind of cage, so you need 2U depth at least.

And above all, if you are using heatsinks, try these out on the bench and measure the heat they produce. A digital thermometer is useful for this, since you can't easily touch hot sinks. Allow for summer weather and the amp being on for several hours. My sinks are 5x8x11 cm per pair of tubes (300b, 46) to give an idea of size.

You need also to give consideration to the gain of the driver. I found 26>46>300b gave enough gain for me, but only with interstages or plate chokes, not resistor loads. You won't easily get over a mu of 8 or so with a DHT that has low enough plate resistance to drive an interstage transformer. The 10Y and 46 have slightly higher gain which is useful, and so does 112A. With 31 and 71a you are down to about 3. I use a plate choke with the 26 because I want the sound and gain of this tube, and it likes inductance of 200H and over. Use a LL1660 in 4.5:1 stepdown and then a step up in the amp and you have a neat solution which allows you to use a TVC. Clever stuff, but it does push the cost up. Since this is a "statement" kind of build, it's worth considering some expensive iron. On the cheap end I use 126C interstage and 156c plate chokes (x2) from Hammond, and these are actually really good and not so far behind Lundahl. Japanese boutique iron is another story!

As you can see, an all-DHT amp is a different proposition and needs to be thought through in terms of real estate and gain and what goes where. A good case for breadboarding.

Andy
 
Last edited:
With a tighter budget and reduced space on my chassis, I ended with two 6SL7 in parallel with CCS loaded and capacitor coupled. Output 45 with Rod Coleman's boards which are superb and fixed bias @ 34mA and Va=300V. I'm planning to move to a 6N7 driver which could fit in my current chassis as per Thomas' suggestion and also add a 26 preamplifier as per Andy's design
So far de results are fantastic, and the 45 sounds superb. Worth trying it yourself as is very difficult to describe in words the natural and rounded sound
Cheers,
Ale
 
Thanks for the inputs Thomas. i will put them onto my consideration. which of course drive a cost a bit up. but i guess would be worth it.

what the disadvantage of using input transformer 1:8 to 1:4 if there is any?

hi Andy
wow that really put me into perspective. its a bit late that i get your excellent idea on putting pieces together into one chassis :( i have already build a preamp with step down LOT. but it open up an option to do like yours for the 2 stage SET amp :) i plan to use 2 top plates for each stage. so trying out different driver tube will be easier in the future. superb. ill be doing diy chassis as well. hence thicker aluminum top plate could act as sink for Rod's Regs. though it will be ugly compare to the finished on the shelf chassis.

ic. guess hi mu DHT is out of question if want to drive the interstage.

Andy, i am interested with #46 as driver but i am totally blind on using dual grid. where do you connect the 1st grid? to the plate? do you have any link to datasheet that has tube 46 graph connected in triode?

if i want to do autobias on #46. Vg=-24v with I=10ma. would give me Rcathode of 2K4? please correct me if wrong.

you are using auto bias with 2 pcs of 12v 23A battery? i am kinda blind on auto bias. please kindly enlighten me. or link on how to do it.

i am interested on learning the filament bias as you did on your 26 too thanks

Thanks so much for the encouragement Ale :) this gonna be long and fun project
 
Last edited:
For 46 in triode connect pins 2 and 4. For voltages see my schematic which would also work into a 126C Hammond. When you use a battery on the grid you need to be sure that this stage is isolated from the preceding one by a capacitor or transformer. If you connect it to solid state components you could do damage. Some basic electronics here!
 
For 46 in triode connect pins 2 and 4. For voltages see my schematic which would also work into a 126C Hammond. When you use a battery on the grid you need to be sure that this stage is isolated from the preceding one by a capacitor or transformer. If you connect it to solid state components you could do damage. Some basic electronics here!

hi Andy

please pardon my ignorant :) so you connect the negative terminal of 24v battery to the grid resistor of 46 and its positive terminal to ground?

how bout the one on 26? do you have the schematic of the choke input supply with a 18+18v transformer?

thanks in advance
erwin
 
Yes, you put two batteries together in series. You will need two N size battery holders for the batteries. Then connect the negative -24v to the bottom of the 270K grid resistor in my schematic and the positive to ground. Monitor the batteries carefully to make sure they are connected and that you have -24v on the grid before you switch anything on and make sure the battery connections are secure and can't come undone at all. Attached my 26 preamp schematic.

It could be useful to have a copy of Morgan Jones "Valve Amps" to check up on some of this. It doesn't cover DHTs in much detail but it's good on the basics.

Andy
 

Attachments

  • 26pre156C_1.jpg
    26pre156C_1.jpg
    656.6 KB · Views: 2,164
do you have any link to datasheet that has tube 46 graph connected in triode?

Nice tubes; quite similar to the 45 but at a fraction of the cost. With the 5 pin socket you are constraining yourself to only use that tube, though. I run mine warm, at 270V p-k and 30V grid, 28mA. They drive an interstage into a 300B without a problem, even at 20 kHz you can darn near drive the grid positive.

Datasheets attached.
 

Attachments

  • 46.pdf
    148.2 KB · Views: 119
  • 46-2.pdf
    146.2 KB · Views: 141
Hi!

what the disadvantage of using input transformer 1:8 to 1:4 if there is any?

None, except that your amp will only work well with certain preamps which are able to drive it properly.

Cathodefollower outputs or solid state preamps typically sound not so good with input transformers IME. Most input transformers don't like any DC offset at all. But a good preamp has none of that anyways.

Another advantage is that you can wire the input differentially now and can support differential connections to the preamp

Best regards

Thomas
 
....With the 5 pin socket you are constraining yourself to only use that tube, though>>

For the 46 there are UX5 sockets that are the same size as UX4 and Octal, so chassis hole can be the same. Actually it's also easy to use those big ceramic oval sockets from China. They also come in octal versions which I have, but have never used. Rather bulky but OK


Andy
 
Hi guys

i draw up some schematic. any input, criticism, suggestion are really welcome.

especially if there is better operating point?

i do not follow mogliaa suggestion of Va 300v. isnt it that 45 max Va is 275v? but i follow operating point from Andy for #46.

all filaments are dc using ccs.

how bout C1 dan C2? are they necessary?

i am planning on using 7903 as Thomas suggested.

does the gain is a bit too small for driver stage? 8*3.5=28x?
any recommendation of DHT that could drive interstage well that has bigger gain than #46?

thanks in advance
erwin
 

Attachments

  • 46-45SET amp.JPG
    46-45SET amp.JPG
    93.9 KB · Views: 525
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.