Hello,
I've decided to give proper DC a try, and this info is buried beneath 100's of pages threads.
I'm looking for opinions from those who have ever compared the CCS modules to large inductance LCL type filament supplies?
I'm mostly curious in sound quality. After pricing both out, there won't be a terrible cost difference.
Thanks
I've decided to give proper DC a try, and this info is buried beneath 100's of pages threads.
I'm looking for opinions from those who have ever compared the CCS modules to large inductance LCL type filament supplies?
I'm mostly curious in sound quality. After pricing both out, there won't be a terrible cost difference.
Thanks
Thank you Gianluca! I hope the Regulators are still sounding good in the "Water Babies".
Quite a few of my kits are in use for RCA 813 (5A) and 75TL (6.25A), and even in parallel pairs for 833 (10A). The transistors burn about 20W (10W each, 2 transistors) into the heatsink at 5.0A. But with transmitter tubes of this size, a fairly big chassis is needed anyway, I expect - so managing the heat is not too difficult. The regulators are soft-starting, to avoid shocking the filament, or dumping uncontrolled thermal pulses into the pass transistors.
The ability to handle extra-high power is another advantage for fully-discrete regulator design - as well as much better sound than pressing ICs into this kind of service!
For your reference, in my 211 line driver (hey! that's water babies ... Rod, good memory!) I use the small side panels of galaxy cabinet by hifi2000 as heatsinks: one Coleman's module per side. It has been running for ... well, I guess 3 years now? more than that?
In my 801A amp, Coleman's CCS' are attached to a thin metal cover that is supporting also tubes, no problem there too.
It is sometime tricky to set precisely the voltage across the filament as it tends to change a bit over time as heatsink gets warmer but it takes only half an hour of patience and that is it!
I never tried DC heated push pull. I have a VT52 differential amp running on AC and apparently there is no need for DC. A friend of mine is using Coleman's on his AD1 PP amp and it was astonishing. Both amps, mine and his, sound really great so I did not mature a definitive idea about AC and DC on push pulls but in SE amps it is a no way back. Old schematics using bridge+Cap or even bridge+Cap+3 pin current regulators are miles away from being perfect.
In my 801A amp, Coleman's CCS' are attached to a thin metal cover that is supporting also tubes, no problem there too.
It is sometime tricky to set precisely the voltage across the filament as it tends to change a bit over time as heatsink gets warmer but it takes only half an hour of patience and that is it!
I never tried DC heated push pull. I have a VT52 differential amp running on AC and apparently there is no need for DC. A friend of mine is using Coleman's on his AD1 PP amp and it was astonishing. Both amps, mine and his, sound really great so I did not mature a definitive idea about AC and DC on push pulls but in SE amps it is a no way back. Old schematics using bridge+Cap or even bridge+Cap+3 pin current regulators are miles away from being perfect.
Attachments
Hi,Rod
Glad to see you here, I've noticed there's a lot of posts of DIY'ers gushing over your regulators. It's my understanding that they are essentially modeling a high impedance inductor. I am curious because of this how they compare to the real thing. No one has really commented on that, maybe if they did it's burried under 100's of posts somewhere. I'd personally like to try both and compare for myself, but one thing at a time.
What I'm getting so far is that for high current, maybe around 5A and greater, stick to chokes because heat becomes an issue, otherwise the Coleman regulator is the way to go.
Just out of curiosity, what size of chokes have you guys compared in your LCL filter to Regulated?
ex) 20-40mH choke or 200-400mH choke.
I'm curious because I think the factor of 10 might be significant.
Glad to see you here, I've noticed there's a lot of posts of DIY'ers gushing over your regulators. It's my understanding that they are essentially modeling a high impedance inductor. I am curious because of this how they compare to the real thing. No one has really commented on that, maybe if they did it's burried under 100's of posts somewhere. I'd personally like to try both and compare for myself, but one thing at a time.
What I'm getting so far is that for high current, maybe around 5A and greater, stick to chokes because heat becomes an issue, otherwise the Coleman regulator is the way to go.
Just out of curiosity, what size of chokes have you guys compared in your LCL filter to Regulated?
ex) 20-40mH choke or 200-400mH choke.
I'm curious because I think the factor of 10 might be significant.
Yes, the value of the choke shows the problem: 40mH is only 25 Ω at 100Hz, and 400mH is only about 250 Ω, so the filtering effect is not so good even with the higher value range.
Also, the small impedance of the choke exposes the filament's (differential) music signal to the filter capacitor (large electrolytic), which is very undesirable.
Using my Regulator, the current-ripple in the filament can easily be reduced to the microampere level, using only a cheap CRC raw dc supply. With LC input, it can be made essentially zero (not really necessary). The regulators do this without adding any (easily measured) self-noise, because the usual noisy bandgap reference has been deliberately avoided.
The 5.0 to 6.5A versions have all the same advantages, and work well within their ratings (eg: 50V durability, 10W dissipation in a 300W FET etc.), and you can use these with confidence. The 5A version has been used for loudspeaker field-coil applications - regulation with lowest noise.
Also, the small impedance of the choke exposes the filament's (differential) music signal to the filter capacitor (large electrolytic), which is very undesirable.
Using my Regulator, the current-ripple in the filament can easily be reduced to the microampere level, using only a cheap CRC raw dc supply. With LC input, it can be made essentially zero (not really necessary). The regulators do this without adding any (easily measured) self-noise, because the usual noisy bandgap reference has been deliberately avoided.
The 5.0 to 6.5A versions have all the same advantages, and work well within their ratings (eg: 50V durability, 10W dissipation in a 300W FET etc.), and you can use these with confidence. The 5A version has been used for loudspeaker field-coil applications - regulation with lowest noise.
Yes, you can. The raw dc supply must float anyway, and 500V should be no problem for the transformer secondary-to-core withstand-voltage, and a split-bobbin EI should be used anyway (to keep the broadband mains noise out).
The Hammond 266 and other split-bobbin types (Hammond say "dual bobbin") are perfect for this application.
The only other precaution is at the transistor heatsink interface: instead of mounting the TO-220s with screws, clips can be used to ensure HT isolation.
The Laird clips act on the body of the TO-220, and avoid any breakdown risk:
TSC607-ZP - LAIRD TECHNOLOGIES - CLIP, 2 X, TO-220 | Farnell UK
Yes, there's a version out in testing now. It has the same audio-frequency isolating qualities as the filament regulator, and the same ultra-low noise. Voltage durability is 250V, or even higher. I hope to have it ready shortly.
The circuits used for fixed bias (RC filter etc) are certainly not good enough for really high quality work.
I mentioned in post 8 that the impedance of a (practical value of) choke is quite low.
Meanwhile, a purpose-designed filament regulator will have a much higher output impedance, Ks of ohms, at least. Adding a choke to the Regulator output will not increase the impedance notably. In the worst outcome, the choke might act somewhat like a search coil, and sniff out unwanted EM noise fields.
You could use the choke in the Regulator's raw dc supply, though - it can act with the raw dc caps to reduce trafo current peaks (and hence, rms current). Use in an LC filter.
Meanwhile, a purpose-designed filament regulator will have a much higher output impedance, Ks of ohms, at least. Adding a choke to the Regulator output will not increase the impedance notably. In the worst outcome, the choke might act somewhat like a search coil, and sniff out unwanted EM noise fields.
You could use the choke in the Regulator's raw dc supply, though - it can act with the raw dc caps to reduce trafo current peaks (and hence, rms current). Use in an LC filter.
I was thinking about an inductors ability to store energy, and started to look into gyrators (since I like to try to understand what I'm doing). There seems to be a practical/theoretical discrepancy regarding gyrators and energy storage (maybe you could help clarify). They really do their job well it seems at impedance inversion, but maybe don't store energy the way an inductor does. I thought best of both worlds would be to put an inductor after the your CCS. Also maybe in this case reactance isn't quite the right word I'm looking for.
I'm thinking it might be a way to use less than 100's mH inductors and sort of turbo charge the regulator at the same time. Have some energy storage real close by the filament.
Anyway, is it worth a try?
I'm thinking it might be a way to use less than 100's mH inductors and sort of turbo charge the regulator at the same time. Have some energy storage real close by the filament.
Anyway, is it worth a try?
If you have the parts, you can always experiment. But I would not be surprised if the sound were degraded - please have a good listen. The Coleman regulator gives microampere-level ripple, or even lower, and no energy storage is required.
Used on the input side, even a small choke will filter some AM or HF radio stations, if these area problem.
Used on the input side, even a small choke will filter some AM or HF radio stations, if these area problem.
I can hardly follow your idea, scutterflux... why do you need 'energy storage' upfront a filament where you only want high impedance and constant current (and voltage)? 100mH at a few amperes of direct current is not a small thing. I found peace of mind with LC smoothing filter followed by regulators.
my 2C
my 2C
It's just about how it sounds.
I bet it would sound different, some guys have said they prefer the sound of LCL others prefer CCS regulators. Now with this CLC-Coleman-L albeit unnecessary and whether or not you could measure anything substantial other than perhaps a minor and unwarranted increase in impedance and a minuscule difference in current is beside the point. The objective would be to try to get the LCL sound without such a large inductor and an even better impedance. Again the real point is only sound and not for any measurable benefits, just for an A/B comparison.
Besides energy storage upfront might mean smoother faster response, easier on the regulator, who knows? Maybe increased dynamics, or better bass etc... Do you ever try things just to see how they sound?
I bet it would sound different, some guys have said they prefer the sound of LCL others prefer CCS regulators. Now with this CLC-Coleman-L albeit unnecessary and whether or not you could measure anything substantial other than perhaps a minor and unwarranted increase in impedance and a minuscule difference in current is beside the point. The objective would be to try to get the LCL sound without such a large inductor and an even better impedance. Again the real point is only sound and not for any measurable benefits, just for an A/B comparison.
Besides energy storage upfront might mean smoother faster response, easier on the regulator, who knows? Maybe increased dynamics, or better bass etc... Do you ever try things just to see how they sound?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.