Hi Rod,
What about using the Keratherm Red transistor insulators (available at diyAudio Store) as an alternative to micas and thermal grease...have you tried that approach?
Best, Robert
Not answering for Rod - but the transistor that needs to be heatsinked here is fully enclosed - so no insulator is needed. Then you are down to just thermal grease vs keratherm.
***************************
Progress is made, if slowly. Lots of metal work, drilling and tapping tonight. Most things are now bolted down to the chassis and I'm getting close to starting wiring. In terms of layout, the transformers and heater raw supply are on the left on a tray that can lift out or sit on springs/sorbothane. Next then are the filament regs, then gyrator boards, then valves on a plate. The distance from gyrator to valves is about 30-40mm. Filament regs to valves is a bit more - probably 75-90mm. All AC is tucked away into corners, and twisted so hopefully my layout is good enough to be quiet. All other wiring is shielded with the shield tied to chassis earth. Circuit ground goes to earth at one point near the IEC module via back to back diodes//10r//10nf cap. I have a bit of thinking to do on optimising grounding, but I think I will use GND on each gyrator board as a mini star, and these will go back to main circuit ground star on the HV board (CRCRC, ground at the last cap terminal). Filament supply grounds then will only connect at the bottom of the (filament) bias resistor to this point as well on each gyrator board.
***************************
Progress is made, if slowly. Lots of metal work, drilling and tapping tonight. Most things are now bolted down to the chassis and I'm getting close to starting wiring. In terms of layout, the transformers and heater raw supply are on the left on a tray that can lift out or sit on springs/sorbothane. Next then are the filament regs, then gyrator boards, then valves on a plate. The distance from gyrator to valves is about 30-40mm. Filament regs to valves is a bit more - probably 75-90mm. All AC is tucked away into corners, and twisted so hopefully my layout is good enough to be quiet. All other wiring is shielded with the shield tied to chassis earth. Circuit ground goes to earth at one point near the IEC module via back to back diodes//10r//10nf cap. I have a bit of thinking to do on optimising grounding, but I think I will use GND on each gyrator board as a mini star, and these will go back to main circuit ground star on the HV board (CRCRC, ground at the last cap terminal). Filament supply grounds then will only connect at the bottom of the (filament) bias resistor to this point as well on each gyrator board.
Last edited:
Looking good, Woodturner.
Would you mind sharing details of your HV supply? What are your CRC values? How about secondary voltage on the transformer?
I understand that the gyrator needs at least 30V above the intended 4P1L plate voltage to operate properly. I am trying to work backwards based on voltage drops I need to account for in the gyrator and also Ale's cap multiplier board.
I was also thinking of using C3D2060F SIC diodes for HT rectification as I have sleeves of them for diode filament bias. The concern is not to exceed the 600V rating. I came across the value of 2.8xsecondary VAC, double of normal solid state rectification.
The question is what is a good VA rating and secondary for prototype testing of a bacic HT CRC supply and eventually Ale's HT boards used with inductors as he has done in some builds.
Is 100VA 185-215VAC a suitable starting point? I don't see any reason to start with more than 300VDC at the first stage of the filter. My biggest issue with Duncan PSUD is knowing I've chosen the right rectifier part.
Would you mind sharing details of your HV supply? What are your CRC values? How about secondary voltage on the transformer?
I understand that the gyrator needs at least 30V above the intended 4P1L plate voltage to operate properly. I am trying to work backwards based on voltage drops I need to account for in the gyrator and also Ale's cap multiplier board.
I was also thinking of using C3D2060F SIC diodes for HT rectification as I have sleeves of them for diode filament bias. The concern is not to exceed the 600V rating. I came across the value of 2.8xsecondary VAC, double of normal solid state rectification.
The question is what is a good VA rating and secondary for prototype testing of a bacic HT CRC supply and eventually Ale's HT boards used with inductors as he has done in some builds.
Is 100VA 185-215VAC a suitable starting point? I don't see any reason to start with more than 300VDC at the first stage of the filter. My biggest issue with Duncan PSUD is knowing I've chosen the right rectifier part.
Better to use a 1200V Schottky. I use these.
https://uk.rs-online.com/web/p/schottky-diodes-rectifiers/8099036
I can't tell you how it will work yet, but I used 600v sic diodes, 150uf caps and 470r at each stage. The transformer secondary is 230v 150ma because that's what I had to hand.... Now I'm wondering if these aren't high enough!
I plan on testing this using some power resistors, burning either 30mA (screen mode) and 60ma (normal anode mode) and seeing the voltage I get. From other builds, the gyrator handles higher voltage input pretty well once you're sensible.
I plan on testing this using some power resistors, burning either 30mA (screen mode) and 60ma (normal anode mode) and seeing the voltage I get. From other builds, the gyrator handles higher voltage input pretty well once you're sensible.
It's true ... if you use it with very low anode swing.
If swing greater, the headroom is
(anode voltage+swing peek voltage) + gyrator headroom.
If you have 170V static anode voltage and swing peek 10V, the B+ must be minimum
170+10+30 (gyrator headroom)= 210V
Maybe worth trying if you are working with 3A-7A filaments; otherwise micas and thermal grease are good. The micas that I supply with the kits are best quality Fischer Elektronik 50µm thickness.
Work continues - the preamp is basically working, but grounding needs to be optimised as its not nearly quiet enough. But it works.
A question (possibly for @mogliaa) I have the pre setup as per the gen 4, screen mode (diagram from Ale's site below). I note that the filament voltage is very low indeed- 1.2V parallel. Why is this? Combat microphonics? And FWIW, the pair of valves I have in there are pretty microphonic - on a par with 01a I would say.
Set up this way, I have ~140V at the "anode", 10mA current, -8.5V grid bias (10r resistor + 3 SIC diodes).
A question (possibly for @mogliaa) I have the pre setup as per the gen 4, screen mode (diagram from Ale's site below). I note that the filament voltage is very low indeed- 1.2V parallel. Why is this? Combat microphonics? And FWIW, the pair of valves I have in there are pretty microphonic - on a par with 01a I would say.
Set up this way, I have ~140V at the "anode", 10mA current, -8.5V grid bias (10r resistor + 3 SIC diodes).
Did you calibrate you Coleman regs for starved filaments at 550mA vs 600mA? There is filament current from regulator plus current from plate to cathode.
I used an online load calculator and was thinking of running mine 140-150. I know I questioned that value before but can't remember right now how I squared it.
I used an online load calculator and was thinking of running mine 140-150. I know I questioned that value before but can't remember right now how I squared it.
Be aware that the 4P1L sounds best at 25mA or more. It really makes a difference. It can sound thin below that. For 10mA I'd rather use a 2P29L which is less fussy about current. Try the 4P1L at 25mA or 30mA and you should hear a difference. This isn't always a practical operating point, which is why I stopped using the 4P1L.
@andyjevans - but this was using screen as anode, hence the lower current.
I did some digging through Ale's website, and I found in his blog comments that he said these were just measured values he got, not design targets as such. So maybe that clears that one up.
Rod - is there a handy way (other than a series resistor) to measure filament current with the regulator? I know I could add in a 1r resistor, but if there is a way already, then great!
I did some digging through Ale's website, and I found in his blog comments that he said these were just measured values he got, not design targets as such. So maybe that clears that one up.
Rod - is there a handy way (other than a series resistor) to measure filament current with the regulator? I know I could add in a 1r resistor, but if there is a way already, then great!
Ok, so grounding I think is sorted out, certainly it's now pretty quiet. The valves are microphonic though, at least as much as 01a and maybe worse.
Some figures:
Filament starved at 550mA
Screen as anode
Anode current 15mA
Anode voltage 150V
Filament bias with grid at -8.8V to K
Circuit as per Ales gen 4 with screen posted above
Star grounding
R-core tx for filaments, EI for HV
B+ 220V
It probably has more gain than I need so I might use a resistor divider at the output to drop the output a bit.
Some figures:
Filament starved at 550mA
Screen as anode
Anode current 15mA
Anode voltage 150V
Filament bias with grid at -8.8V to K
Circuit as per Ales gen 4 with screen posted above
Star grounding
R-core tx for filaments, EI for HV
B+ 220V
It probably has more gain than I need so I might use a resistor divider at the output to drop the output a bit.
Why not use divider at input?
Using after output possible lose low/er/ output impedance, which is requiring to load problematic interconnects.
@euro21 - I'm shamelessly following the path others have trod in front of me:
https://www.bartola.co.uk/valves/2020/12/29/01a-low-gain-dht-preamp/
I got a bit of a listen here this evening, but could only spend a short while having a listen. While its all there, has depth and width, it doesn't have that same midrange that the 01a has. I'm wondering if the microphonics are causing some of this, while you don't notice them as you're listening, if the music stops suddenly, you can hear the slight chime for a second or so. If that is happening all the time then it may well be contributing to what I'm hearing. I still have some hum as well, its low enough, but there all the same.
So that's the "fix" list. On the positives, the soundstage is really very good indeed and if I hadn't had the 01a as my daily driver I would be ecstatic with the 4P1L. The extension top and bottom is very good, and dynamics are great, really its very very good once I can fix up the few niggles. No doubt they are down to me!
Even further good news, the set of 2P29L are on the way to me now as well, so I must go read up on those as well.
BTW, is there an english translated datasheet PDF for the 4P1L anywhere?
Fran
https://www.bartola.co.uk/valves/2020/12/29/01a-low-gain-dht-preamp/
I got a bit of a listen here this evening, but could only spend a short while having a listen. While its all there, has depth and width, it doesn't have that same midrange that the 01a has. I'm wondering if the microphonics are causing some of this, while you don't notice them as you're listening, if the music stops suddenly, you can hear the slight chime for a second or so. If that is happening all the time then it may well be contributing to what I'm hearing. I still have some hum as well, its low enough, but there all the same.
So that's the "fix" list. On the positives, the soundstage is really very good indeed and if I hadn't had the 01a as my daily driver I would be ecstatic with the 4P1L. The extension top and bottom is very good, and dynamics are great, really its very very good once I can fix up the few niggles. No doubt they are down to me!
Even further good news, the set of 2P29L are on the way to me now as well, so I must go read up on those as well.
BTW, is there an english translated datasheet PDF for the 4P1L anywhere?
Fran
Yes - the virtues of the 4P1L are dynamics and extension top and bottom. The midrange can be a bit hollow, which is where it loses out to other DHTs, even the 2P29L. The midrange fills in a bit more at 30mA.
Thanks for the pic @jasonb84!
@andyjevans - I need to go back and listen again. I thought as I got to the end it was sounding "better" and I wondered was it just that everything had warmed up nicely at that stage (about 30mins). I also thought microphonics had settled a little bit by then. Anyway, I think I'll get more time tomorrow to try this for longer.
You know the reason I had 4P1L originally is that years ago I bought a headphone amp from china - it was called a little dot 2 I think and it had DHT tubes (I had no idea what DHT was back then), 4P1L, and I think the driver tube was something like 1s4 or something. I sold it to a friend but kept a set of the 4P1L in the stash from back then. That was probably the mid noughties.
@andyjevans - I need to go back and listen again. I thought as I got to the end it was sounding "better" and I wondered was it just that everything had warmed up nicely at that stage (about 30mins). I also thought microphonics had settled a little bit by then. Anyway, I think I'll get more time tomorrow to try this for longer.
You know the reason I had 4P1L originally is that years ago I bought a headphone amp from china - it was called a little dot 2 I think and it had DHT tubes (I had no idea what DHT was back then), 4P1L, and I think the driver tube was something like 1s4 or something. I sold it to a friend but kept a set of the 4P1L in the stash from back then. That was probably the mid noughties.