kmaier,
Aha! You taught me something new.
I wasn't aware that tube selection for the best specs meant so much to minimize hum.
I'll continue to search for ways to reduce hum on my pedestrian tubes, but at least I won't try to make a silk purse out of a sow's ear.
Frank
Aha! You taught me something new.
I wasn't aware that tube selection for the best specs meant so much to minimize hum.
I'll continue to search for ways to reduce hum on my pedestrian tubes, but at least I won't try to make a silk purse out of a sow's ear.
Frank
"Low hum" is indeed relative- a few millivolts at the speaker terminals is often good enough- but a tenth of a millivolt is certainly possible. Separate grounds (star) for each stage helps, separate filtering/bypassing for the B+ on each stage helps, well-filtered B+ to begin with, certainly shielded input cables, even the "phase" of the heater connections (try reversing the two heater pins).
Likely the best results are obtained when the signal's star ground is NOT coincident with the ac mains "green wire" ground- use a single wire to connect the star to the earth ground. (just my opinion/experience. Each case is different).
Likely the best results are obtained when the signal's star ground is NOT coincident with the ac mains "green wire" ground- use a single wire to connect the star to the earth ground. (just my opinion/experience. Each case is different).
Go with the Sound Card/PC Analyzer.
Comments from a Newbie, I fully support what Tom said about a Sound Card/PC based Spectrum Analyzer. This is fairly inexpensive set up and a solid investment.
I was using one but wasn't sure if it was really telling the truth. I picked up a HP 334A Distortion Analyzer at a garage sale. Now I believe my PC Analyzer.
When I'm using the HP if feel I'm working in the dark. Fortunately is has an output that you can connect a Spectrum Analyzer to and see what's left when you remove the fundamental.
To be honest, now that I have the HP I rarely use it.
Another really cool feature is the FR sweep. This a very quick way of testing FR from 20hz to 20khz for range and flatness.
Comments from a Newbie, I fully support what Tom said about a Sound Card/PC based Spectrum Analyzer. This is fairly inexpensive set up and a solid investment.
I was using one but wasn't sure if it was really telling the truth. I picked up a HP 334A Distortion Analyzer at a garage sale. Now I believe my PC Analyzer.
When I'm using the HP if feel I'm working in the dark. Fortunately is has an output that you can connect a Spectrum Analyzer to and see what's left when you remove the fundamental.
To be honest, now that I have the HP I rarely use it.
Another really cool feature is the FR sweep. This a very quick way of testing FR from 20hz to 20khz for range and flatness.
While one can claim output noise (or signal-to-noise ratio) to be relative, it's effects are not. A poor s/n can easily affect low-level detail and smear it. You really need to understand the system as a whole before declaring an acceptable level of hum and/or noise. Achieving an acceptable s/n goes beyond grounding techniques... a proper design and attention to detail on layout are important. You also need to consider isolation of signal, power and filament lines, quality components, magnetic coupling and the actual vacuum tubes employed as all of these can heavily influence the overall performance.
For testing, I have a HP 334A and several 400FL RMS meters (all calibrated). I also have multiple digital audio cards with multiple (pro-level) 24-bit external A/D and D/A converters, plus an assortment of typical PC sound cards and of course laptops with built-in sound cards. I've used PC sound cards with software before, but I still prefer using my HP measurement gear (and a good scope) when spec'ing out tube gear... and I've never felt in the dark using it. It's simple, reliable, very consistent and has a wide (signal) input range, something that PC sounds cards don't. As a result, you need to manage the I/O signals with the PC sound card to match levels with the equipment under test.... which requires additional external equipment which brings it's own specs into the picture. As a result, you end up with measurements that are more relative than absolute. I'm not saying it's not useful... but you need to understand the limitations of the test equipment you're using and how it interfaces to the equipment under test.... compensating for this can be the larger challenge.
Regards, KM
For testing, I have a HP 334A and several 400FL RMS meters (all calibrated). I also have multiple digital audio cards with multiple (pro-level) 24-bit external A/D and D/A converters, plus an assortment of typical PC sound cards and of course laptops with built-in sound cards. I've used PC sound cards with software before, but I still prefer using my HP measurement gear (and a good scope) when spec'ing out tube gear... and I've never felt in the dark using it. It's simple, reliable, very consistent and has a wide (signal) input range, something that PC sounds cards don't. As a result, you need to manage the I/O signals with the PC sound card to match levels with the equipment under test.... which requires additional external equipment which brings it's own specs into the picture. As a result, you end up with measurements that are more relative than absolute. I'm not saying it's not useful... but you need to understand the limitations of the test equipment you're using and how it interfaces to the equipment under test.... compensating for this can be the larger challenge.
Regards, KM
Gentlemen,
I have the hum down to .3mV. I added virtual CTs to ground on the filament transformers & padded the humpots so they aren't knife edge sensitive.
My next step is to raise the filament voltages by connecting them to the B+.
How much should I raise the preamp (7N7, aka 6SN7)? Should I also raise the driver (7N7) & the power tube(2A3) ?
Frank
I have the hum down to .3mV. I added virtual CTs to ground on the filament transformers & padded the humpots so they aren't knife edge sensitive.
My next step is to raise the filament voltages by connecting them to the B+.
How much should I raise the preamp (7N7, aka 6SN7)? Should I also raise the driver (7N7) & the power tube(2A3) ?
Frank
I couldn't agree more. I use stand-alone test equipment for exactly those reasons.
A computer sound card is just one way to get going with some actual measurements without having to break the bank. It's something that's readily available for most people. I just personally prefer the stand-alone gear. I'd much rather spin a dial than click with a mouse.
~Tom
I read that as 0.3 mV or 300 uV. Is that correct? If that's the case, it's pretty damn good. Congrats.
Elevating the heater in an indirectly heated tube (7N7/6SN7) will violate the heater-cathode max voltage, likely result in arc-over, and destroy the tube and/or amplifier. Your best option is likely to tie the heater "center tap" to the cathode.
For directly heated triodes (2A3), the filament is the cathode. How the filament is connected to ground (or negative bias supply) determines the plate current. Don't mess with it unless you know exactly what you're doing.
~Tom
tomch,
OMG! I 'only' have 0.6mV hum. The ".3mV" I typed was the target I was aiming for.
I'm glad you told me about the indirect heated tubes. I was planning on elevating their voltage today. You saved me from a catastrophe, thanks. I won't mess with elevating the 2A3 voltage, either.
I'll connect the 7N7 heater CT to the cathode .
Frank
OMG! I 'only' have 0.6mV hum. The ".3mV" I typed was the target I was aiming for.
I'm glad you told me about the indirect heated tubes. I was planning on elevating their voltage today. You saved me from a catastrophe, thanks. I won't mess with elevating the 2A3 voltage, either.
I'll connect the 7N7 heater CT to the cathode .
Frank
2A3 hum levels
Dobias, et al,
Getting low hum can be difficult with AC filaments but not impossible. Again, my personal spec is 80dB down from 1-watt RMS output (8- or 16-ohm does not matter... hence a dB spec, not voltage).
As your circuit has cathode biasing on the 2A3, you may want to try a split-balance arrangement. I do this on all of my 45 amps and it works well on the 2A3, but caveats apply (see next paragraph). I use the filament transformer center-tap (real, not virtual) as a fixed DC balance, meaning you tie the center-tap to the cathode biasing resistor (which is NOT bypassed). I use a 100 ohm linear pot across the filament and pad the pot with a pair of 12-ohm resistors. The pot wiper has the bypass cap connected there. As a result, you have a fixed DC path (transformer center-tap and cathode bias resistor) and an adjustable AC path (padded pot and bypass cap). Give it a try... it may work well for you, but....
There are quite a few variations of the 2A3 (assumed all NOS types)... the original single-plate version... which had a more complex series/parallel filament arrangement and must be powered with DC to get anywhere near an acceptable hum level. A revised (and much cheaper to produce) twin-triode version. These can be full dual triodes in parallel or can be a pair of filaments and grids with a single plate structure forming two independent triodes in parallel. The latter versions usually perform pretty well with the split balance approach described above. Also note that you could reverse it and have a fixed AC balance and adjustable DC balance or use dual padded pots and have independently adjustable AC and DC balance (have done this as well on a 45 SET design).
I would also look to obtain some select NOS 2A3 types/brands if you can find them. I have a fairly good and diverse selection of 2A3 tubes, both single plate and the various dual-plate versions noted above. Brands include RCA, Sylvania, Ken-Rad, Philco, Zenith and maybe another. In RCA I have both the gray- and black-plate versions. I'll cut this short (sans the reasoning, etc.) but from a personal preference view, my choices are:
1- Sylvania (with the coil-spring filament supports)
2- Sylvania (with the fishing-rod filament supports)
3- Ken-Rad (true twin-triodes with blacked-out glass)
4- RCA Black-Plate
5- RCA Gray-Plate
Note that Philco/Zenith brands (at least the ones I have) are identical to the Sylvanias and were undoubtedly sourced from Sylvania. As a result they fit into 1 or 2 above.
Regards, KM
Dobias, et al,
Getting low hum can be difficult with AC filaments but not impossible. Again, my personal spec is 80dB down from 1-watt RMS output (8- or 16-ohm does not matter... hence a dB spec, not voltage).
As your circuit has cathode biasing on the 2A3, you may want to try a split-balance arrangement. I do this on all of my 45 amps and it works well on the 2A3, but caveats apply (see next paragraph). I use the filament transformer center-tap (real, not virtual) as a fixed DC balance, meaning you tie the center-tap to the cathode biasing resistor (which is NOT bypassed). I use a 100 ohm linear pot across the filament and pad the pot with a pair of 12-ohm resistors. The pot wiper has the bypass cap connected there. As a result, you have a fixed DC path (transformer center-tap and cathode bias resistor) and an adjustable AC path (padded pot and bypass cap). Give it a try... it may work well for you, but....
There are quite a few variations of the 2A3 (assumed all NOS types)... the original single-plate version... which had a more complex series/parallel filament arrangement and must be powered with DC to get anywhere near an acceptable hum level. A revised (and much cheaper to produce) twin-triode version. These can be full dual triodes in parallel or can be a pair of filaments and grids with a single plate structure forming two independent triodes in parallel. The latter versions usually perform pretty well with the split balance approach described above. Also note that you could reverse it and have a fixed AC balance and adjustable DC balance or use dual padded pots and have independently adjustable AC and DC balance (have done this as well on a 45 SET design).
I would also look to obtain some select NOS 2A3 types/brands if you can find them. I have a fairly good and diverse selection of 2A3 tubes, both single plate and the various dual-plate versions noted above. Brands include RCA, Sylvania, Ken-Rad, Philco, Zenith and maybe another. In RCA I have both the gray- and black-plate versions. I'll cut this short (sans the reasoning, etc.) but from a personal preference view, my choices are:
1- Sylvania (with the coil-spring filament supports)
2- Sylvania (with the fishing-rod filament supports)
3- Ken-Rad (true twin-triodes with blacked-out glass)
4- RCA Black-Plate
5- RCA Gray-Plate
Note that Philco/Zenith brands (at least the ones I have) are identical to the Sylvanias and were undoubtedly sourced from Sylvania. As a result they fit into 1 or 2 above.
Regards, KM
kmaier,
The tubes I have are Shuguang 2A3C. Once I get the bugs out of this amp, I'll look for your recommended tubes.
The humpots are 100 ohm & padded with 10 ohm resistors. Their wiper is already connected to a 47uf bypassed 1.0k resistor to ground. Please elaborate , or sketch, the connection change.
Since my tubes are single plate I understand you suggest ,"reverse it and have a fixed AC balance and adjustable DC balance or use dual padded pots and have independently adjustable AC and DC balance."
I'm willing to follow your lead, either way, but ask that you provide sketches for me to follow.
I hope I'm not asking too much, but, with the help I'm getting through this minefield, I don't want to back into any more problems.
Frank
The tubes I have are Shuguang 2A3C. Once I get the bugs out of this amp, I'll look for your recommended tubes.
The humpots are 100 ohm & padded with 10 ohm resistors. Their wiper is already connected to a 47uf bypassed 1.0k resistor to ground. Please elaborate , or sketch, the connection change.
Since my tubes are single plate I understand you suggest ,"reverse it and have a fixed AC balance and adjustable DC balance or use dual padded pots and have independently adjustable AC and DC balance."
I'm willing to follow your lead, either way, but ask that you provide sketches for me to follow.
I hope I'm not asking too much, but, with the help I'm getting through this minefield, I don't want to back into any more problems.
Frank
split balance
Frank, attached is a low-cost 45 amp I designed a while back... take a look at the cathode circuitry.
Regards, KM
NOTE: don't reverse it... start with fixed DC balance and adjustable AC balance.
Frank, attached is a low-cost 45 amp I designed a while back... take a look at the cathode circuitry.
Regards, KM
NOTE: don't reverse it... start with fixed DC balance and adjustable AC balance.
kmaier,
I can follow the schematic with your written instructions until the center tap through where the resistor connects to the first cap of the CLC filter..... I have an LC (15H & 4uf) to dual LC's dual (4.5H & 200uF) filter.
Do I connect the resistor to the negative of the first (4uf) cap, or should I add another cap ahead of the first choke?
Frank
I can follow the schematic with your written instructions until the center tap through where the resistor connects to the first cap of the CLC filter..... I have an LC (15H & 4uf) to dual LC's dual (4.5H & 200uF) filter.
Do I connect the resistor to the negative of the first (4uf) cap, or should I add another cap ahead of the first choke?
Frank
Frank,
The cathode bias resistor connects to the filament transformer center-tap at one end and the other end goes to ground... look closer at the schematic.
Regards, KM
The cathode bias resistor connects to the filament transformer center-tap at one end and the other end goes to ground... look closer at the schematic.
Regards, KM
kmaier,
OK, now I see the 1.5k 10w resistor connected to the CT and connecting to a line that goes to the PT CT & to a ground. The crossing lines fooled me. May I ask why the 15k doesn't have it's own ground symbol? Does this indicate the need for a common star ground for all of the items in the PS?
Frank
OK, now I see the 1.5k 10w resistor connected to the CT and connecting to a line that goes to the PT CT & to a ground. The crossing lines fooled me. May I ask why the 15k doesn't have it's own ground symbol? Does this indicate the need for a common star ground for all of the items in the PS?
Frank
Frank,
Simply the way I draw up my schematics, if the lines connect on paper they are connected in circuit. If they are not suppose to connect, there is an arc that shows the line crossing but not connected. With any design, proper grounding techniques apply. For all of my SET designs I use a single large gauge ground bus that is isolated from chassis ground.
Regards, KM
Simply the way I draw up my schematics, if the lines connect on paper they are connected in circuit. If they are not suppose to connect, there is an arc that shows the line crossing but not connected. With any design, proper grounding techniques apply. For all of my SET designs I use a single large gauge ground bus that is isolated from chassis ground.
Regards, KM
kmaier,
Now that you describe it, I was aware of your drawing method but I 'assumed' that connecting wires needed a dot. As the kids say, "my bad!"
It's a struggle for this old dog, but I'm learning, thanks to guys like you that have infinite patience.
I, too, used the large gauge ground bus.
Frank
Now that you describe it, I was aware of your drawing method but I 'assumed' that connecting wires needed a dot. As the kids say, "my bad!"
It's a struggle for this old dog, but I'm learning, thanks to guys like you that have infinite patience.
I, too, used the large gauge ground bus.
Frank
kmaier,
I tried a quick fix, with the components I have, without any improvement.
You have: a real CT, 1.5k 10w resistor, 20uf cap., 12R padding resistors
I have: a virtual CT, 1.0k 10w resistor, 47uf cap., 10R padding resistors
Unless I need to change to match your parts exactly, I may as well accept the
status quo.
Thank you for your efforts. My rat's nest amp may be beyond further improvement.
At least I'll know enough to have a much larger chassis next time.
When I get the ambition, I'll move the filament transformers & rectifiers outside the chassis so I can clean up the wiring some more.
Best Regards,
Frank
I tried a quick fix, with the components I have, without any improvement.
You have: a real CT, 1.5k 10w resistor, 20uf cap., 12R padding resistors
I have: a virtual CT, 1.0k 10w resistor, 47uf cap., 10R padding resistors
Unless I need to change to match your parts exactly, I may as well accept the
status quo.
Thank you for your efforts. My rat's nest amp may be beyond further improvement.
At least I'll know enough to have a much larger chassis next time.
When I get the ambition, I'll move the filament transformers & rectifiers outside the chassis so I can clean up the wiring some more.
Best Regards,
Frank
- Status
- Not open for further replies.