Sy,
- why did you put a 4uF/400V in the "B+" ? What if I put there a 200V capacitor?
- what voltage rating should be the output capacitor?
- what voltage rating should be the input capacitor, if I would like to omitt the transformer ?
- what specific transistors do you recomment for the current source?
- If I have a separate winding on my transormfer for heating the tube, do I need that 2x120 resistors paralell near the current source?
Thanks!
P.S.: I tried to edit my previous post, but it took me too much, and the 30minutes editing period expired, I couldnt edit that post
- why did you put a 4uF/400V in the "B+" ? What if I put there a 200V capacitor?
- what voltage rating should be the output capacitor?
- what voltage rating should be the input capacitor, if I would like to omitt the transformer ?
- what specific transistors do you recomment for the current source?
- If I have a separate winding on my transormfer for heating the tube, do I need that 2x120 resistors paralell near the current source?
Thanks!
P.S.: I tried to edit my previous post, but it took me too much, and the 30minutes editing period expired, I couldnt edit that post
You're right, I forgot to spec the transistors. I used NTE199. There's nothing special about that choice- it was based on what rattled out of the coffee can first when I shook it. Anything with a BVceo of 50V or better, an ft better than 10MHz, and a beta of 100 or greater will work. The NTE is a 50V device, ft of 90MHz, beta of 400, and costs under a buck.
The plate bypass cap can absolutely be a 200V device. I used a 400V cap because that's what I had on hand. If I were to rationalize, I'd say that it was to give me more flexibility on tube choice. For this tube, the extra voltage rating isn't needed.
The output cap will never see more than 15V. It can have any voltage rating you like. I had some Rifa caps with a 160V rating that I used, but a 63V Wima will be perfect there.
Likewise, any input cap will never see high voltage.
The heaters here are running on DC. If you use your winding (presumably 6V?), rectify and smooth it, chances are that you'll still need the dropping resistors, but you'll have to recalculate the value.
The plate bypass cap can absolutely be a 200V device. I used a 400V cap because that's what I had on hand. If I were to rationalize, I'd say that it was to give me more flexibility on tube choice. For this tube, the extra voltage rating isn't needed.
The output cap will never see more than 15V. It can have any voltage rating you like. I had some Rifa caps with a 160V rating that I used, but a 63V Wima will be perfect there.
Likewise, any input cap will never see high voltage.
The heaters here are running on DC. If you use your winding (presumably 6V?), rectify and smooth it, chances are that you'll still need the dropping resistors, but you'll have to recalculate the value.
Very interesting. I also believed that galvanic isolation must be beneficial. I had my TVC hooked with an isolated ground for a while but in some source configurations this added a noticeable hum. Before switching to a common ground i listened very carefully and could not hear any difference at all when the input and output grounds were shorted. Ok, the hum disappeared when shorting. More recently i have been using the TX102 in an autoformer configuration in which there is obviously a single earth connection with great success.
These experiences make me question any inherent advantages in keeping earth loops galvanically isolated. It makes good sense in theory but success seems dependent upon the specific curcumstances.
SY said:
Like some people are religious about cables
)Thanks your advises, I will choose MKT MKS4.
It's 10uF/100V (theresn't smaller voltage rated in this capacity)
What about that resistor, at the heating-part?
What do you think about putting BC546B into the constant current generator?
What values do you recommend into the input R-C ?
I just used some off-the-shelf rectangular power resistors for dropping the heater voltage. In theory, since the ECC88 has parallel heaters, it should be fed by a voltage source, not a current source. I'm not convinced that it's a big deal either way, so a resistor splits the difference in a Solomonic fashion.
Input RC? My personal inclination would be to use 0.1uF and 1M, inserted between the wiper and the grid. That choice is based on careful consideration of the position of the low frequency rolloff traded off with grid leakage current and my religious belief in not using large coupling caps whenever possible. And dictated mostly by what I have in my coffee can, which at the moment is flush with 0.1uF caps.
Input RC? My personal inclination would be to use 0.1uF and 1M, inserted between the wiper and the grid. That choice is based on careful consideration of the position of the low frequency rolloff traded off with grid leakage current and my religious belief in not using large coupling caps whenever possible. And dictated mostly by what I have in my coffee can, which at the moment is flush with 0.1uF caps.
Don't you think it would be better, if that's where a system problem is, to add the input transformers to the power amps? We're diyers. We can do that, we've got the technology.
IIRC, input transformers do a better job of isolation. But if that's not right, I'm confident you'll gleefully correct me.
IIRC, input transformers do a better job of isolation. But if that's not right, I'm confident you'll gleefully correct me.
SY said:
Well personally I prefer to use input transformers when it's practical. But trying to incorporate an input transformer into an existing piece of equipment often isn't terribly practical compared to including say an output transformer in a ground-up project such as this one.
Anyway, I was just addressing your comment that you didn't see that an output transformer would bring anything to the table.
se
Milan, I'll give you a sneak peek- the +/-12V supply will power an opamp in the next version. I hate that output capacitor and will show how to be rid of it.
The heaters are actually a positive here- they have a very constant draw and help load down the 12 volt regulators. This was a trick I picked up from Dave Berning.
The heaters are actually a positive here- they have a very constant draw and help load down the 12 volt regulators. This was a trick I picked up from Dave Berning.
SY said:
Hi,
input - transformer
active element - tube
ccs - bipolar
output/servo - opamp...
And to round things up properly, +B PS should be realized with a vertical Fet and a chipamp (properly snubberized and upgraded to classA) should be used as a tube heater. Now you're set to go
Regards,
Milan
P.S. What about a selen +B rectifier?
jeff mai said:
That's a stretch! If it were me, I'd be soldering up some BUF03s or whatever is available like that these days.
moamps said:
Hi,
input - transformer
active element - tube
ccs - bipolar
output/servo - opamp...
And to round things up properly, +B PS should be realized with a vertical Fet and a chipamp (properly snubberized and upgraded to classA) should be used as a tube heater. Now you're set to go
Regards,
Milan
P.S. What about a selen +B rectifier?
Bite your tongue!
Yeah, there's still a technology or two that needs to get stuck in there. I point out that there's already a MOSFET, just not in the signal path. And you didn't mention the LED. I could isolate the -12V rail to the CCS with a BUF03...
The only "heresy" I see here is: if the source can drive a 10K transformer, and the load is 10K, why do we need a tube at all?
OK, you are looking to drive the sometimes-large RF-suppression caps in power amp inputs. And empty-out coffee-cans of spare parts. And do a little education on the side. Kewl.
> These experiences make me question any inherent advantages in keeping earth loops galvanically isolated. It makes good sense in theory but success seems dependent upon the specific curcumstances.
Very much so.
When all your gear is on one shelf, one power outlet, it should be possible to get scrupulously clean with simple direct unbalanced interfaces.
SY apparently has a source with roof antenna plus lightning-ground. Murphy's Law tells us that "ground" at the antenna is never the same as "ground" at the hi-fi shelf. Hence he has discernable dirt, and a transformer is a time-proven cure.
When you get into bigger systems, diff-inputs become vital. My concert hall has different power feeds for stage and booth. I get 0.1V to 2V AC between "ground" downstairs and upstairs. Unbalanced connections drown in hum. My PA system has two line inputs: an active diff-amp for sources inside the booth, and a good transformer input for feeds outside the booth. The iron ensures that ground-difference will never crud-up the sound, even if grounding fails (see Murphy, op cit).
Similar issues in recording studios bigger than a bedroom usually lead to balanced inputs everywhere. Very high level outputs should be balanced to reduce crosstalk, though in fact many studio "balanced outputs" are not truly balanced.
> In theory, since the ECC88 has parallel heaters, it should be fed by a voltage source, not a current source. I'm not convinced that it's a big deal either way
AFAICT, a heater is a heater. It is just a resistor that gets hot. Before the 1930s, heater resistance was not tightly specified: apply rated voltage and it would work fine, but the current was not always what the book said. Then came series-string radios and especially TV sets. If a tube heater's resistance was way off from nominal, it would over- or under-heat, and work poorly. So they got more careful about heater resistance, and about the change of resistance from cold to hot. At first only for the 5 tubes used in radios, but TV sets used about every tube in the book, so they made all heaters to tight specs for voltage, current, and warm-up.
Since utility power is basically voltage-source, it is usually most efficient to feed heaters a voltage source. But whatever turns you on.
> input transformers do a better job of isolation. But if that's not right, I'm confident you'll gleefully correct me.
If you can't float both ends (expensive and lossy), then putting the iron on the input usually gives better isolation and will generally be cheaper. You know the input sensitivity and thus the transformer maximum level. Outputs may be asked to drive a variety of loads, and you have to buy iron for the largest expected level.
> Would it work well with a low imedance attenuator (100 ohms DCR, 1200 ohms @ 20Hz) on the output?
Read SY's inital posts. He scaled for a 10K load, and explained his calculations. You wanna drive a 1.2K load as generously as he drives a 10K load, you will need 8 times the cathode current, and 4 to 8 tubes or a much larger tube than he proposes. SY's 10K plan is generous enough that it "would" drive 1.2K "OK", in the sense that it would not clip. But THD would be much higher than SY is aiming for.
OK, you are looking to drive the sometimes-large RF-suppression caps in power amp inputs. And empty-out coffee-cans of spare parts. And do a little education on the side. Kewl.
> These experiences make me question any inherent advantages in keeping earth loops galvanically isolated. It makes good sense in theory but success seems dependent upon the specific curcumstances.
Very much so.
When all your gear is on one shelf, one power outlet, it should be possible to get scrupulously clean with simple direct unbalanced interfaces.
SY apparently has a source with roof antenna plus lightning-ground. Murphy's Law tells us that "ground" at the antenna is never the same as "ground" at the hi-fi shelf. Hence he has discernable dirt, and a transformer is a time-proven cure.
When you get into bigger systems, diff-inputs become vital. My concert hall has different power feeds for stage and booth. I get 0.1V to 2V AC between "ground" downstairs and upstairs. Unbalanced connections drown in hum. My PA system has two line inputs: an active diff-amp for sources inside the booth, and a good transformer input for feeds outside the booth. The iron ensures that ground-difference will never crud-up the sound, even if grounding fails (see Murphy, op cit).
Similar issues in recording studios bigger than a bedroom usually lead to balanced inputs everywhere. Very high level outputs should be balanced to reduce crosstalk, though in fact many studio "balanced outputs" are not truly balanced.
> In theory, since the ECC88 has parallel heaters, it should be fed by a voltage source, not a current source. I'm not convinced that it's a big deal either way
AFAICT, a heater is a heater. It is just a resistor that gets hot. Before the 1930s, heater resistance was not tightly specified: apply rated voltage and it would work fine, but the current was not always what the book said. Then came series-string radios and especially TV sets. If a tube heater's resistance was way off from nominal, it would over- or under-heat, and work poorly. So they got more careful about heater resistance, and about the change of resistance from cold to hot. At first only for the 5 tubes used in radios, but TV sets used about every tube in the book, so they made all heaters to tight specs for voltage, current, and warm-up.
Since utility power is basically voltage-source, it is usually most efficient to feed heaters a voltage source. But whatever turns you on.
> input transformers do a better job of isolation. But if that's not right, I'm confident you'll gleefully correct me.
If you can't float both ends (expensive and lossy), then putting the iron on the input usually gives better isolation and will generally be cheaper. You know the input sensitivity and thus the transformer maximum level. Outputs may be asked to drive a variety of loads, and you have to buy iron for the largest expected level.
> Would it work well with a low imedance attenuator (100 ohms DCR, 1200 ohms @ 20Hz) on the output?
Read SY's inital posts. He scaled for a 10K load, and explained his calculations. You wanna drive a 1.2K load as generously as he drives a 10K load, you will need 8 times the cathode current, and 4 to 8 tubes or a much larger tube than he proposes. SY's 10K plan is generous enough that it "would" drive 1.2K "OK", in the sense that it would not clip. But THD would be much higher than SY is aiming for.
Did you say regulator? Since I got drafted into child care today, I'm running behind on drawing and write-up for the power supply. As a quick preview, here's a drawing of the B+ regulator (and yes, there is some definite logic to using a simple two transistor job).
Attachments
