Then you might find another improvement using the Salas HV shunt regulator.
http://www.diyaudio.com/forums/power-supplies/134801-simplistic-mosfet-hv-shunt-regs-88.html
http://www.diyaudio.com/forums/power-supplies/134801-simplistic-mosfet-hv-shunt-regs-88.html
Thanks. I'll try this too. I was reading the same thread just before I saw your post.
G’ Day All,
It has been a while since anyone posted updates on their 26 project, I suspect many are busy hunting Arcturus 126 tubes on ebay, see the prices the last few went for. Don’t listen to Coolzero, the blue globe 126 are not that special.
I managed to finally finish my filament power supply, I’m using a 60mH choke input (followed by 32000uf & common mode choke) all sealed up in a steel chassis. The output voltage looked really nice on a scope as a neat sine wave. I wired up the filament with a ccs and switch that allows me to use the 26 and 01A. The net results of the choke input filament supply was a more defined bass, I’m still yet to add 0.22uF snubber caps and lower the Q as per Morgan Jones’ suggestion.
I found that shielding the plate chokes in aluminium boxes also helped and that if shielding the tubes with Al or Cu, holding it down with rubber “O” rings can help a lot. The pearl tube coolers also work well to reduce microphonics.
I have a question in regard to the Russian FT-3 teflon caps, did anyone ground the metal can/shield or is it better to just remove it?
Has anyone that has experimented with other output caps, have results to share?
Cheers,
Rich
It has been a while since anyone posted updates on their 26 project, I suspect many are busy hunting Arcturus 126 tubes on ebay, see the prices the last few went for. Don’t listen to Coolzero, the blue globe 126 are not that special.
I managed to finally finish my filament power supply, I’m using a 60mH choke input (followed by 32000uf & common mode choke) all sealed up in a steel chassis. The output voltage looked really nice on a scope as a neat sine wave. I wired up the filament with a ccs and switch that allows me to use the 26 and 01A. The net results of the choke input filament supply was a more defined bass, I’m still yet to add 0.22uF snubber caps and lower the Q as per Morgan Jones’ suggestion.
I found that shielding the plate chokes in aluminium boxes also helped and that if shielding the tubes with Al or Cu, holding it down with rubber “O” rings can help a lot. The pearl tube coolers also work well to reduce microphonics.
I have a question in regard to the Russian FT-3 teflon caps, did anyone ground the metal can/shield or is it better to just remove it?
Has anyone that has experimented with other output caps, have results to share?
Cheers,
Rich
Attachments
Last edited:
Hi - nice work there. I have a couple of 26 preamps to experiment with. I've settled on choke loaded operation, with FT-2 output caps - the smaller 200v versions. Ixys cascode was a bit sharp, Holco resistor a shade muffled and weaker bass, though nice sound. Plate chokes just about right. I'm satisfied enough by now to be able to put my energies into my 300b SET. That's now sounding pretty good. VT-67 into 31 into 300b. Separate power supply with a nice old Cossor 53KU (GZ33) rectifier. All DHT system - bliss! I have to say, building that lot gave me a few grey hairs - 8 DC filament supplies for starters and two power supplies - separate one for the preamp. Plenty of stuff to go wrong there. But it really sounds nice, as of course it should!!!!
andy
andy
Hi Andy,
Thanks for the reply. Did you notice a difference with the different voltage rated caps?
After hearing what a 26 preamp can do I’m moving towards an all DHT set up as well. I have all the parts for a 31 interstage 46 amp, just need to start the build, the 26 pre should provide just enough gain. I also reckon separate filament supplies are the way to go, more over choke input supplies.
Rich
Thanks for the reply. Did you notice a difference with the different voltage rated caps?
After hearing what a 26 preamp can do I’m moving towards an all DHT set up as well. I have all the parts for a 31 interstage 46 amp, just need to start the build, the 26 pre should provide just enough gain. I also reckon separate filament supplies are the way to go, more over choke input supplies.
Rich
Hi - 31 and 46 are both very nice, but you're dropping the gain a bit. Why not use something like 01A in the input of the amp to get a bit more gain - or a VT-67 is nice too, or indeed another 26? 10Y is a lovely driver - possibly the best. 46 is nice and has a nice tone and neutral sound but 10Y is more vivid and dynamic. 31 is good and saves on the filament requirements
I found the FT-2 and 3 pretty similar and just a fraction ahead of the K72 versions. They are all pretty good.
I'm lazy with filaments - just use LM1084 instead of chokes. But you do need heatsinks so you can't escape metal parts!
andy
I found the FT-2 and 3 pretty similar and just a fraction ahead of the K72 versions. They are all pretty good.
I'm lazy with filaments - just use LM1084 instead of chokes. But you do need heatsinks so you can't escape metal parts!
andy
Andy,
I’ll give the 10Y a shot as an input, I was saving them as an output tube, but at 1.6 watts there wouldn’t be much of a difference compared to the 46.
In your previous posts on an all DHT 300B amp, you’ve indicated that the 300B does not sound slow and like a typical 300B, could you elaborate on the sound compared to the 10Y or 801A as output tubes (if you’ve tried them).
I’d also like to hear your thoughts/comments on an 01A vs 26 pre, driving that 300b amp
Rich
I’ll give the 10Y a shot as an input, I was saving them as an output tube, but at 1.6 watts there wouldn’t be much of a difference compared to the 46.
In your previous posts on an all DHT 300B amp, you’ve indicated that the 300B does not sound slow and like a typical 300B, could you elaborate on the sound compared to the 10Y or 801A as output tubes (if you’ve tried them).
I’d also like to hear your thoughts/comments on an 01A vs 26 pre, driving that 300b amp
Rich
I think as a preamp tube I'd go 10Y, 26, 01A in that order, but it's pretty tight. I never used the 10y as an output - I like the 300b sound when it's driven by a 10Y so that would be my ideal. If I'm feeling cheap and lazy I use VT-67 and 31 just because I have a lot of small 6v transformers and 01A and 10Y really should have 9v secondaries for use with a LM1086/5/4. The 01A may work with a 6v secondary - it's a bit marginal. It's certainly a better sound than the VT-67 and no more expensive these days. But it takes a lot to beat a 26 and they're cheap enough.
andy
andy
I think a modular approach works great. For my amps (and indeed preamps) I use a variety of different top plates. All 275mm deep, and 100mm, 70mm or 50mm wide. That enables me to use 2U rackmount parts, which end up 420mm wide allowing for the joins in the plates. OPTs go on 100mm, output tubes on 70mm and the rest on 50mm or 100mm as desired. You basically just need to use UX4 sockets unless you use 46 or other 5 pin tubes.
I use a separate power supply. In fact I'm thinking of two smaller boxes for the PSU - one for filaments and one for the HT - that allows for more experiments by using different filament supplies. It's been a long time designing all this but beginning to get good results!
andy
I use a separate power supply. In fact I'm thinking of two smaller boxes for the PSU - one for filaments and one for the HT - that allows for more experiments by using different filament supplies. It's been a long time designing all this but beginning to get good results!
andy
In terms of DC filament supplies, batteries don't hum but I leave those for those who have the patience to build and use them. I don't have super sensitive speakers, so I'm not the best person to answer this. For me it isn't the hum, it's the sound quality. I think a big capacitor and any kind of current source work fine for hum, and sound good enough for me. If I'm lazy I just use a single LM1084/5/6 on the right size heatsink. If I'm going for the max I use Rod Coleman's design (in this thread) which to me is as good as it gets. I haven't used choke input but others swear by that. I put the schottky diodes and the first cap in the outboard PSU and the second cap and current source in the signal chassis and use 4 pin XLRs for umbilicals. I use a 4 pin Speakon for the HT since it's shrouded and rated for 300v.
andy
andy
Just a note of revised operating points for me 26 pre, which is choke loaded using two Hammond 156C chokes (150H at 8mA) back to back, giving 6.6k resistance, and uses Russian teflon FT-2 caps out.
HT 375v
Dropper resistor 4.2k 14watts
Glow tubes: Two VT105 giving 210v out
Plate voltage - 175v
Cathode resistor 1.9K wirewound
Vk = 10.5
Current = 5.5mA
Sounds like I'm preferring the globe 26 in this setup. All very nice!
andy
HT 375v
Dropper resistor 4.2k 14watts
Glow tubes: Two VT105 giving 210v out
Plate voltage - 175v
Cathode resistor 1.9K wirewound
Vk = 10.5
Current = 5.5mA
Sounds like I'm preferring the globe 26 in this setup. All very nice!
andy
Hi - don't have a circuit diagram right now but could scare one up. It's pretty simple as described
HT is any way you want it - I try to use a tube rectifier, choke input or at least a small cap like 1uF, then motor run caps after the choke - for the glow tubes at 210v you probably need about 300v so they strike properly. You can use the smaller 7 pin types or the octals - the octals give you a bit more current capability, though you only need a max of 14mA. I don't know if there's any difference in sound. Right now I have 375v before the glow tubes. I use a 4 pin Speakon as umbilical from the PSU to signal chassis.
Then in the signal chassis:
47uF motor run cap
Dropper resistor 4.2k 14watts
Glow tubes: Two VT105 giving 210v out
Plate chokes are two 156C hammond, bolted bottom to bottom and wired out of phase, i.e. 1A to 2B, 1B to 2A. Here's the digikey reference.
CHOKE DC FILTER 150H 8MA - 156C
Plate voltage - 175v
Cathode resistor 1.9K wirewound - 1.8K probably better.
Vk = 10.5
Current = 5.5mA
The above values are approximate and vary with the tubes used.
Filaments are Schottky diodes and 2,200uF cap in PSU box, 4 pin XLR umbilical. In signal chassis, 15000uF cap and LM1084 wired as current source with output from the adjust pin and 1.25 resistor (7 watt or more) between adj pin (1) and out pin (2). The LM1084 sits on a heatsink 3" by 2" by1".
Russian FT-2 caps 0.1uF out and short lead into amp (tube).
That should do it.
andy
HT is any way you want it - I try to use a tube rectifier, choke input or at least a small cap like 1uF, then motor run caps after the choke - for the glow tubes at 210v you probably need about 300v so they strike properly. You can use the smaller 7 pin types or the octals - the octals give you a bit more current capability, though you only need a max of 14mA. I don't know if there's any difference in sound. Right now I have 375v before the glow tubes. I use a 4 pin Speakon as umbilical from the PSU to signal chassis.
Then in the signal chassis:
47uF motor run cap
Dropper resistor 4.2k 14watts
Glow tubes: Two VT105 giving 210v out
Plate chokes are two 156C hammond, bolted bottom to bottom and wired out of phase, i.e. 1A to 2B, 1B to 2A. Here's the digikey reference.
CHOKE DC FILTER 150H 8MA - 156C
Plate voltage - 175v
Cathode resistor 1.9K wirewound - 1.8K probably better.
Vk = 10.5
Current = 5.5mA
The above values are approximate and vary with the tubes used.
Filaments are Schottky diodes and 2,200uF cap in PSU box, 4 pin XLR umbilical. In signal chassis, 15000uF cap and LM1084 wired as current source with output from the adjust pin and 1.25 resistor (7 watt or more) between adj pin (1) and out pin (2). The LM1084 sits on a heatsink 3" by 2" by1".
Russian FT-2 caps 0.1uF out and short lead into amp (tube).
That should do it.
andy
Hi Regal,
No matter what DC power supply one uses, it is still hard to attenuate the hum. I started off with a simple cap filter, followed by voltage regulator then CCS like Andy’s setup. I then upgraded to a larger transformer (12v 5A, just because I had it lying around) and choke input filter 60mH 2A, followed by common mode choke, see pics in post 303. I followed the advice here and had it all sealed up in steel enclosure. Only by shielding the tubes themselves did I get it quiet enough. I should also point out my speakers are rather efficient at 97db. I haven’t tried Rod’s or Salas’ regulators yet. I also posted a link to some tests on choke input filament supplies in post 289
My B+ supply is very similar to Andy’s, but I’m also using chokes in there and a VR 105 and VR 90 for a B+ of around 188V.
Andy,
You make it all sound so easy. Do you have your cathode resistor bypassed in that setup?
I’d really like to see some pictures of your creations.
Rich
No matter what DC power supply one uses, it is still hard to attenuate the hum. I started off with a simple cap filter, followed by voltage regulator then CCS like Andy’s setup. I then upgraded to a larger transformer (12v 5A, just because I had it lying around) and choke input filter 60mH 2A, followed by common mode choke, see pics in post 303. I followed the advice here and had it all sealed up in steel enclosure. Only by shielding the tubes themselves did I get it quiet enough. I should also point out my speakers are rather efficient at 97db. I haven’t tried Rod’s or Salas’ regulators yet. I also posted a link to some tests on choke input filament supplies in post 289
My B+ supply is very similar to Andy’s, but I’m also using chokes in there and a VR 105 and VR 90 for a B+ of around 188V.
Andy,
You make it all sound so easy. Do you have your cathode resistor bypassed in that setup?
I’d really like to see some pictures of your creations.
Rich
Rod,
With regard to the LM4041 adj used as the current set in your 26 filament schematic, you wrote
This part appears to have been discontinued and is very difficult to source.
The LM385 adj http://www.national.com/ds/LM/LM185-ADJ.pdf
also appears to be suitable. The only significant difference I can see is that Vmax is 5.3V instead of 10V, which I don't think would preclude its use in this application.
Can you see any issues ?
With regard to the LM4041 adj used as the current set in your 26 filament schematic, you wrote
This part appears to have been discontinued and is very difficult to source.
The LM385 adj http://www.national.com/ds/LM/LM185-ADJ.pdf
also appears to be suitable. The only significant difference I can see is that Vmax is 5.3V instead of 10V, which I don't think would preclude its use in this application.
Can you see any issues ?
The LM385 seems to meet the spec just fine.
But with a low voltage durability, and tiny supply current demands, the feed-in resistor R6 would be better increased from 3.3K to 50K, 100K or even 220K. Probably best to try these values and listen, starting with the highest. High values would react with the gate capacitance to roll-off noise, which is certainly beneficial. A little mica capacitor of 470p .. 3.3nF [R6-LM385 to GND] could also be added to emphasise the effect.
But with a low voltage durability, and tiny supply current demands, the feed-in resistor R6 would be better increased from 3.3K to 50K, 100K or even 220K. Probably best to try these values and listen, starting with the highest. High values would react with the gate capacitance to roll-off noise, which is certainly beneficial. A little mica capacitor of 470p .. 3.3nF [R6-LM385 to GND] could also be added to emphasise the effect.