NewYorkDave
Well-known member
Did you include the "extended" requirements provided by PRR? I defer to him since I'm certainly no transformer expert, just an end-user like most of yuz.
Small Line Amp (Tube)
- Thread starter NewYorkDave
- Start date
Help Support GroupDIY Audio Forum:
tk@halmi
Well-known member
[quote author="NewYorkDave"]Did you include the "extended" requirements provided by PRR?[/quote]
Yes.
Yes.
> a good 15k:650 trafo, check this:
Looks good except: I think we are looking for a transformer that will take some DC. My Portuguese is poor, but I don't see DC specs or the magic phrase "air gap" (which seems to be the same in Portuguese as in english).
They do have air-gap output iron, but all wound to speaker impedance.
I didn't see prices. Do they do export shipping?
Do they do "specials"? A good SE 2A3 or 6L6 speaker tranny, but secondary wound to 600 ohms, could be sweet and not hard to compute. Use 8 or 9 times as many turns of wire 1/3rd the overall diameter.
Nickle is very interesting, but sometimes we want a "cheap" sound. If the music is good, a little M6 won't hurt.
Looks good except: I think we are looking for a transformer that will take some DC. My Portuguese is poor, but I don't see DC specs or the magic phrase "air gap" (which seems to be the same in Portuguese as in english).
They do have air-gap output iron, but all wound to speaker impedance.
I didn't see prices. Do they do export shipping?
Do they do "specials"? A good SE 2A3 or 6L6 speaker tranny, but secondary wound to 600 ohms, could be sweet and not hard to compute. Use 8 or 9 times as many turns of wire 1/3rd the overall diameter.
Nickle is very interesting, but sometimes we want a "cheap" sound. If the music is good, a little M6 won't hurt.
rafafredd
Well-known member
No No No. No airgap on this transformer. Guilhereme is currently working on gaped versions.
NewYorkDave
Well-known member
By the way, that custom Edcor would work well for an RCA BC-2B preamp, if they hit all the specs you gave 'em.
skipwave
Well-known member
[quote author="tk@halmi"]Yes, it will have a Ni core. It is not cheap although you know how much more competitors ask for similar things. It could be used with the One-Bottle micpre also. Thanks for the specs Dave![/quote]
Tamas - Did you happen to get a price quote with the M6 core, instead of the Ni?
Thanks,
Brian
Tamas - Did you happen to get a price quote with the M6 core, instead of the Ni?
Thanks,
Brian
edcorusa
Active member
[quote author="NewYorkDave"]That's expensive by Edcor's price standards. Are they using a nickel core or somethin'?[/quote]
I agree that it is expensive. EDCOR have tried to keep its prices down. The problem is that nickel is over four times the price of steel and because it is a softer metal and has to be hand stacked. The wastage going through a stacking machine is also too expansive. This is one of the reasons that EDCOR stopped using nickel steel at one point. Now with more and more people asking EDCOR to build transformers with nickel, we have opened the doors again.
I agree that it is expensive. EDCOR have tried to keep its prices down. The problem is that nickel is over four times the price of steel and because it is a softer metal and has to be hand stacked. The wastage going through a stacking machine is also too expansive. This is one of the reasons that EDCOR stopped using nickel steel at one point. Now with more and more people asking EDCOR to build transformers with nickel, we have opened the doors again.
tk@halmi
Well-known member
[quote author="skipwave"]Tamas - Did you happen to get a price quote with the M6 core, instead of the Ni?[/quote]
The feedback I got indicated that Dave's specs cannot be satisfied with M6 lams. With the M6 core the dillema is: how many dB loss are you willing to take in addition to the 14dB step down due to DCR of the added turns? I am sure it can be done on an M6 core for half the price of the Ni core.
If you are not picky about a little dB loss you could give it a try and use Dave's specs, omit the DCR part and specify the M6 lams. Their engineer is good at figuring out the rest.
Cheers,
Tamas
The feedback I got indicated that Dave's specs cannot be satisfied with M6 lams. With the M6 core the dillema is: how many dB loss are you willing to take in addition to the 14dB step down due to DCR of the added turns? I am sure it can be done on an M6 core for half the price of the Ni core.
If you are not picky about a little dB loss you could give it a try and use Dave's specs, omit the DCR part and specify the M6 lams. Their engineer is good at figuring out the rest.
Cheers,
Tamas
tk@halmi
Well-known member
The Edcor plate transformer has shipped. It will be here in a week.
tk@halmi
Well-known member
The part number is NXSE1-600-15k. Testing is on my ever growing list...
skipwave
Well-known member
Thanks for the update, Tamas. I'm looking forward to hearing your impressions.
NewYorkDave
Well-known member
Note: if replacing the output plate resistor and coupling cap with a gapped 15K:600 transformer like Tamas', you need to increase the value of the output tube cathode resistor. These values should get you in the ballpark:
1K (for One-Bottle pre)
510 ohms (for small line amp posted in this thread).
In either case, the output tube plate resistance will be a good match for the 15K primary. Current draw of either circuit will remain about the same as the original versions. Output tube plate dissipation will increase, but will still be within safe limits, and maximum output power will increase by a fair amount. The feedback resistor can be increased slightly to compensate for the ~2dB loss of net gain from using a 5:1 transformer ratio (compared to the 4:1 specified originally).
1K (for One-Bottle pre)
510 ohms (for small line amp posted in this thread).
In either case, the output tube plate resistance will be a good match for the 15K primary. Current draw of either circuit will remain about the same as the original versions. Output tube plate dissipation will increase, but will still be within safe limits, and maximum output power will increase by a fair amount. The feedback resistor can be increased slightly to compensate for the ~2dB loss of net gain from using a 5:1 transformer ratio (compared to the 4:1 specified originally).
edanderson
Well-known member
i built up the output tube half of this lineamp today, basically from the .022uf cap into the grid of V2 on, with a UTC A-25 as a shunt fed output transformer (the A-25 is only rated for 8ma DC). without any termination on the secondary of the transformer, it makes about 18db of gain, and there is quite a spike around 50KHz. loading the output with 600 ohms brings the overall gain down to about 14db, and rounds off the top end rather severely. 1kohm is better, 15db of gain, flat with no ringing and a gentler rolloff around 50KHz. in all configurations there is about a 2db bump in the bass centered around 20Hz. with a bit of value twiddling it seems like the frequency response should be quite good.
the problem that i had was that THD is quite high -- steady around 11% from 80KHz all the way down to 10Hz. that the level of distortion is nearly ruler flat across the whole spectrum suggests that i did something wrong. will this stage not work on its own without negative feedback?
ed
the problem that i had was that THD is quite high -- steady around 11% from 80KHz all the way down to 10Hz. that the level of distortion is nearly ruler flat across the whole spectrum suggests that i did something wrong. will this stage not work on its own without negative feedback?
ed
NewYorkDave
Well-known member
You've introduced too many variables at the same time, and left out too much information. Please don't make us go hunting for data on that transformer--at least post some basic specs like impedance ratio. The distortion figure you give doesn't mean much to us because you don't specify input/output levels and load impedance. You don't tell us voltages. You also don't specify what you're trying to achieve by using the output stage on its own.
Anyway, as a reality check, try it cap-coupled into a resistive load and then take it from there.
Anyway, as a reality check, try it cap-coupled into a resistive load and then take it from there.
edanderson
Well-known member
UTC A-25: 15k to 600 ohm
input / output levels: -10dbu in, between +4dbu and +8dbu out depending on output loading
source impedance: i tried 50, 150, and 600 ohm source impedances; no major effect on THD
load impedance: i tried (virtually) no load, a 600 ohm resistor, and a 1k resistor, as reported in my post
voltages: B+ was 300v; i will check the voltages at plate and cathode and post those
what i was trying to achieve: paul stamler mentioned near the beginning of the thread that just the output tube would make a decent output driver by itself if you didn't need the voltage gain of the first tube. that is what i was going for.
ed
input / output levels: -10dbu in, between +4dbu and +8dbu out depending on output loading
source impedance: i tried 50, 150, and 600 ohm source impedances; no major effect on THD
load impedance: i tried (virtually) no load, a 600 ohm resistor, and a 1k resistor, as reported in my post
voltages: B+ was 300v; i will check the voltages at plate and cathode and post those
what i was trying to achieve: paul stamler mentioned near the beginning of the thread that just the output tube would make a decent output driver by itself if you didn't need the voltage gain of the first tube. that is what i was going for.
ed
tk@halmi
Well-known member
[quote author="edanderson"]will this stage not work on its own without negative feedback?[/quote]
Generally, plucking a portion out of an overall design is a tough cookie.
The negative feedback reduces distortion in several ways. One is the correction factor, depending on the magnitude of the feedback. Another is its effect on lowering the output impedance, and driving the transformer better. I would not be surprised if the global feedback brought the distortion down into the vicinity of 1% and improved the frequency transfer function substantially.
As much bad rap as global negative feedback gets, when done correctly it really kicks butt, and far more effective than C37 lacquer.
Generally, plucking a portion out of an overall design is a tough cookie.
The negative feedback reduces distortion in several ways. One is the correction factor, depending on the magnitude of the feedback. Another is its effect on lowering the output impedance, and driving the transformer better. I would not be surprised if the global feedback brought the distortion down into the vicinity of 1% and improved the frequency transfer function substantially.
As much bad rap as global negative feedback gets, when done correctly it really kicks butt, and far more effective than C37 lacquer.
> THD is quite high -- steady around 11% from 80KHz all the way down to 10Hz.
> input / output levels: -10dbu in, between +4dbu and +8dbu out depending on output loading
load impedance: i tried (virtually) no load, a 600 ohm resistor, and a 1k resistor,
If the THD meter is showing 11% at various frequencies, levels, and loads: the meter is sick.
I would expect higher levels with lower loads to show a clear rise of THD.
Do try a 10K series resistor with the source. If THD soars, you are pulling grid current, which is usually a bad plan. You can "get away with it" when you have 50 ohm sources; but if you have a 50 ohm source, why do you need a line driver? Often we have a consumer output, or a pot, or an EQ network, a medium-high-Z source, and that's why we need a line driver.
> input / output levels: -10dbu in, between +4dbu and +8dbu out depending on output loading
load impedance: i tried (virtually) no load, a 600 ohm resistor, and a 1k resistor,
If the THD meter is showing 11% at various frequencies, levels, and loads: the meter is sick.
I would expect higher levels with lower loads to show a clear rise of THD.
Do try a 10K series resistor with the source. If THD soars, you are pulling grid current, which is usually a bad plan. You can "get away with it" when you have 50 ohm sources; but if you have a 50 ohm source, why do you need a line driver? Often we have a consumer output, or a pot, or an EQ network, a medium-high-Z source, and that's why we need a line driver.
edanderson
Well-known member
voltages:
B+ 300V
between the 4.7k and 10k resistors 248V
at the plate 152V
grid -0.025V
cathode 2.5V
[quote author="PRR"]If the THD meter is showing 11% at various frequencies, levels, and loads: the meter is sick. [/quote]
i doubt that is the case. i would more readily fault the meter READER (myself) or the guy who built the circuit (also myself).
i tried putting a 10k resistor in series with the input (signal generator > 10k > 0.022uf > tube grids and 1Gohm to ground) and it actually reduced THD.
with only the analyzer as a load on the transformer secondary, with a -10dbu input signal, and a 600 ohm source, the THD from 100Hz to 10KHz the THD is about 10%. below 100Hz, THD rises to about 35% at 10Hz. above 10KHz it rises to a peak of 45% around 40KHz and then falls rapidly above 50KHz.
when i add a 10k resistor in series with the input to this same configuration, the response below 100Hz is basically the same at 33%. from 100Hz to 10KHz it is about 8%. from 10KHz up, the peak around 40KHz is still there, but the tip is now inverted, so there are two lesser peaks, at about 30KHz and 55KHz.
ed
B+ 300V
between the 4.7k and 10k resistors 248V
at the plate 152V
grid -0.025V
cathode 2.5V
[quote author="PRR"]If the THD meter is showing 11% at various frequencies, levels, and loads: the meter is sick. [/quote]
i doubt that is the case. i would more readily fault the meter READER (myself) or the guy who built the circuit (also myself).
i tried putting a 10k resistor in series with the input (signal generator > 10k > 0.022uf > tube grids and 1Gohm to ground) and it actually reduced THD.
with only the analyzer as a load on the transformer secondary, with a -10dbu input signal, and a 600 ohm source, the THD from 100Hz to 10KHz the THD is about 10%. below 100Hz, THD rises to about 35% at 10Hz. above 10KHz it rises to a peak of 45% around 40KHz and then falls rapidly above 50KHz.
when i add a 10k resistor in series with the input to this same configuration, the response below 100Hz is basically the same at 33%. from 100Hz to 10KHz it is about 8%. from 10KHz up, the peak around 40KHz is still there, but the tip is now inverted, so there are two lesser peaks, at about 30KHz and 55KHz.
ed
NewYorkDave
Well-known member
Did you try eliminating the transformer and cap-coupling to a 10K resistive load? Let's see if the amplifier circuit itself is sick first, before complicating things with the transformer.
Similar threads
