Khmmm....
I don't like roasted resistors. 🙂
"High Temperature Exposure 250 h at + 250 °C for U Characteristic"
Hello Triodos,
The Coleman Regulator controls current, and has a trimmer to adjust current. When connected to a filament, it can be used to adjust the filament voltage to exactly the correct Voltage value. This works perfectly on every kind of DHT I know! I have tried many kinds, and I find that the filament always behaves like a resistor - certainly well enough to adjust it. And I'm sure the many other constructors using the regulator have found it works the same way.
So there is no problem operating on constant current - but you must use a regulator that is very stable both with applied voltage-difference, and temperature. If the circuit is not designed to ignore changes in BOTH of these external effects, then you will find that (for example) temperature effects change the current, and the current-change affects the voltage, which causes more instability in the current. The same regulator problems can make the filament look like a constant current: e.g. the current is adjusted lower, but this makes the power supply voltage rise, which in turn forces the current higher again. The basic circuit in the picture has all these problems - it was just a starting-point for a usable regulator.
Hi Rod,
Thanks for explanation, seems I need to try different (read: better) current source, and I will report back about voltage and current stability.
Thanks for explanation, seems I need to try different (read: better) current source, and I will report back about voltage and current stability.
Hello and thanks for sharing your experience. I found myself puzzled on how a change in components would produce such a noticeable and favorable outcome when the reason for using the regulator is because it does a very good job in isolating the audio signal from the DHT filament power supply. The reason I believe is that it has to do with the frequency response seen for the audio signal from the filament to the regulator. Different components will affect the frequency response. What you have shared is actually good because it has shown components can be rolled to achieve favorable results. I may have this all wrong...
Adrien.
(Currently enjoying V4 regulators which replaced VCCS regulators)
Hi.
Had reconfigured my DC supply, so it now runs much colder. But as you have mention, I'll upgrade to the Mills MRA-12, their will run colder.
now... just checking. can i run left and right channels rod's regulators from the same raw supply?
Hi Gianluca!
With any heating scheme, the L and R channels must not be coupled.
Although the Coleman Regulator buffers both sides of the filament, there may be unwanted current flow, if the Inputs are connected together. This is certain to be a problem, if auto (cathode) bias is used - since the cathode voltage dc levels are sure to be different.
Are you working on something new?
With any heating scheme, the L and R channels must not be coupled.
Although the Coleman Regulator buffers both sides of the filament, there may be unwanted current flow, if the Inputs are connected together. This is certain to be a problem, if auto (cathode) bias is used - since the cathode voltage dc levels are sure to be different.
Are you working on something new?
I knew that ... nothing comes for free.
Rod, I am dismantling my big mono amps 801A-75TL to savage parts and build a 801A-75TL (or 211) amp for stax. Actually I do not own any stax headphone ... in case I sell some stuff I'll consider buying myself a pair. I have no room in my system for those big amps any more.
I'd like to stay <20kg and 1 chassis, so from 4 high voltage supplies I am stepping down to a single shared PSU and I was considering ways to save room, parts and weight for filaments. Heating those bottles is a demanding task as I do maintain choke input filters 'sound better'.
I might share the transformer and attach two bridges and filters+regulators, that's still something.
Rod, I am dismantling my big mono amps 801A-75TL to savage parts and build a 801A-75TL (or 211) amp for stax. Actually I do not own any stax headphone ... in case I sell some stuff I'll consider buying myself a pair. I have no room in my system for those big amps any more.
I'd like to stay <20kg and 1 chassis, so from 4 high voltage supplies I am stepping down to a single shared PSU and I was considering ways to save room, parts and weight for filaments. Heating those bottles is a demanding task as I do maintain choke input filters 'sound better'.
I might share the transformer and attach two bridges and filters+regulators, that's still something.
Hi.
I have learned it but the hard way, but now I'm up running on Rod Coleman ver7 in
both my amps (4P1L,2A3).
I'm on the old version of 4P1L still the 8ohm, and my loftin white with Monolith Summit S-9.
I have reconfigured my AC/DC in my 4P1L, so I now can run on only 7.5 volt AC.
I only have 3.4 volt in difference across the coleman ver 7 board, and I can set the
filament from 1.4-2.0 volt on my 4P1L.
Now running on starved filament on 1.55 volt, and the first I noticed was less microphonic from my 4P1L even when playing quite loud.
Having 12 watt less power from my Trafos means my heat sink can be touched even
after 2 hours of usage.
The clarity have improved even further. I have Oscons/Mills in my boards and as previous said, YOU NEED TO GO FOR MILLS-MRA12 in a 2A3.
In 4P1L Mills-MRA5 is sufficient. I have reused the Mills/Oscon from my ver 5, so I
don't need the burnin time of these components.
Will post later one when I have more time on the ver 7.
Keep up your great work Rod...
/Michael.
I have learned it but the hard way, but now I'm up running on Rod Coleman ver7 in
both my amps (4P1L,2A3).
I'm on the old version of 4P1L still the 8ohm, and my loftin white with Monolith Summit S-9.
I have reconfigured my AC/DC in my 4P1L, so I now can run on only 7.5 volt AC.
I only have 3.4 volt in difference across the coleman ver 7 board, and I can set the
filament from 1.4-2.0 volt on my 4P1L.
Now running on starved filament on 1.55 volt, and the first I noticed was less microphonic from my 4P1L even when playing quite loud.
Having 12 watt less power from my Trafos means my heat sink can be touched even
after 2 hours of usage.
The clarity have improved even further. I have Oscons/Mills in my boards and as previous said, YOU NEED TO GO FOR MILLS-MRA12 in a 2A3.
In 4P1L Mills-MRA5 is sufficient. I have reused the Mills/Oscon from my ver 5, so I
don't need the burnin time of these components.
Will post later one when I have more time on the ver 7.
Keep up your great work Rod...
/Michael.
Hi Rod,
Have you ever given more thought to making a regulator for PP DHT amps? I'm about to start my plans for probably a PP 300B amp, but I am also considering low voltage GM-70, and would like to have a good filament regulator.
Thanks,
John
Have you ever given more thought to making a regulator for PP DHT amps? I'm about to start my plans for probably a PP 300B amp, but I am also considering low voltage GM-70, and would like to have a good filament regulator.
Thanks,
John
Hi John - Yes I did work through the design, but it was a dead-end. The problems involved in combining a PP pair for all applications are not easily overcome.
Simply building PP with individually-regulated heating is the most efficient and best-sounding route.
Thanks, Michael - I am pleased the amps are moving on up!
The OS-CONs in the C2 position are OK, but I would prefer to see Panasonic FM/FR or the original caps at C1. OS-CONs have desirable low-impedance over a wide bandwidth but their leakage current is high and unstable, which may degrade performance, especially at higher temperatures.
Thanks for the feedback!
Hi Rod.
Thanks for your input. I have used the Panasonic FM before and though it's not as detail as the for instance the Elna Silmic 2, over time I like it better.
What I like with the Oscons is the "organic" sound. They are very "open" but they never has this "imponator effect" which only last for a very short
time in my opinion.
Your Nichicon C2 on ver 7, is really a good Caps, hard to beat.
As told before the Mills R1 in my opinion is a "winner" but it will make your board to expensive.
As said before when a new version of your regulators is ready, Denmark is ready too.. 🙂
/michael
Wow. I didn't know the V7 is out now. I just finished assembling 4 boards with label V4.2001-2011 for my 845 amps. I bought them last year and didn't have time to come around to them till now. Parts are well labeled and everything looks nice and easy to assembly. But I wish the board is a bit larger and parts are spaced a bit further apart for the guys with big hands like me.
Rod, the question is what's the difference between V4 2001-2011 and the V7 boards?
I was looking into my stash of old heatsinks but didn't find the right one for the job. I almost put in the new order but I remembered I have a large amount of CPU heatsinks and I found some old Pentium III heatsinks looks perfect. Even the holes aligned perfectly for the board.
Regards,
James
Rod, the question is what's the difference between V4 2001-2011 and the V7 boards?
I was looking into my stash of old heatsinks but didn't find the right one for the job. I almost put in the new order but I remembered I have a large amount of CPU heatsinks and I found some old Pentium III heatsinks looks perfect. Even the holes aligned perfectly for the board.
Regards,
James
Hi James,
The new V7 incorporates all the refinements I could come up with over the 3 years the V4 was in service. The best updates: the V7's dynamic impedance looking out of the filament is now higher and flatter (vs freq.) - and is the reason for the reports here of improved sound. The V7 also works with a lower headroom of supply voltage - about 1V lower - so they can run cooler.
But the V4 will still give excellent performance - enjoy those DHT amps!
Not ready to rip the v4 out of my amps or preamp any time soon. But just knowing there's a v7 makes my inner tweaker start to whisper to me ...
Rod, what max ripple can V7 run with? Assuming I need to heat up a 300B, I'll want 6VDC min + how much ripple?
Hi Gianluca - For 300B: with raw dc ripple at 150mV peak-to-peak, expect a ripple-current in the output to be a few µA (translates to <25µV in the resistance of the 300B filament). Across the audio band, beyond the mains ripple: say 150Hz-20kHz the noise floor is much lower still.
This is much quieter than necessary for a power stage, so you can allow raw dc ripple to double or triple without problems: but the minimum voltage of the input waveform (bottom of the saw-tooth) must stay above 3.0V or so above the desired filament voltage), in all cases. The 3.0V headroom for a 300B is an absolute minimum - please allow for raw dc tolerances. With the V4, 4.0V headroom is required.
This is much quieter than necessary for a power stage, so you can allow raw dc ripple to double or triple without problems: but the minimum voltage of the input waveform (bottom of the saw-tooth) must stay above 3.0V or so above the desired filament voltage), in all cases. The 3.0V headroom for a 300B is an absolute minimum - please allow for raw dc tolerances. With the V4, 4.0V headroom is required.