hi all
i've been reading these pages for a long time but this is my first post.
I have searched the archives (and the rest of the net) extensively and i have not been able to find a solution.
I am an experienced tube-audio hobbyist (I have scratch-built 25- 30 tube audio pieces in the past few years). I recently built my first tube phono pre.
I used the standard RCA receiving-tube manual circuit married to the standard RCA low impedance 'input amp' as an output stage. The unit sounds GREAT, i mean, really really great, but it has pretty bad RF in the first stage. WEBE 108 in stamford CT. This station, BTW, often gets into tube studio microphones in the area.
I use a Grado green series (AKA the $60 one) cartridge. 56k imp. I use a 56k resistor from input grid to ground in the circuit.
I do NOT use a grid stopper resistor (the circuit did not call for one).
Any help would be greatly appreciated. I have several solid-state cheapo phone pre amps that work totally fine with this same deck and cart., so i assume this problem is correctable?
here are links to the original circuit if you care to check it out:
http://www.bonavolta.ch/hobby/en/audio/rca.htm
as i said before, i am using a 56k input load resistor instead of the 47k indicated here.
c/
i've been reading these pages for a long time but this is my first post.
I have searched the archives (and the rest of the net) extensively and i have not been able to find a solution.
I am an experienced tube-audio hobbyist (I have scratch-built 25- 30 tube audio pieces in the past few years). I recently built my first tube phono pre.
I used the standard RCA receiving-tube manual circuit married to the standard RCA low impedance 'input amp' as an output stage. The unit sounds GREAT, i mean, really really great, but it has pretty bad RF in the first stage. WEBE 108 in stamford CT. This station, BTW, often gets into tube studio microphones in the area.
I use a Grado green series (AKA the $60 one) cartridge. 56k imp. I use a 56k resistor from input grid to ground in the circuit.
I do NOT use a grid stopper resistor (the circuit did not call for one).
Any help would be greatly appreciated. I have several solid-state cheapo phone pre amps that work totally fine with this same deck and cart., so i assume this problem is correctable?
here are links to the original circuit if you care to check it out:
http://www.bonavolta.ch/hobby/en/audio/rca.htm
as i said before, i am using a 56k input load resistor instead of the 47k indicated here.
c/
I also recomend using a grid stopper.....
Normally the RIAA filter would filter the RF, but in your case the RF is entering the grid and being "detected"...demodulated to audio band...then your phono stage is amplifying the unwanted audio signal....
I also suggest using Hi-End resistors in that location.....first you want to hit the loading resistor of say 47K or 56K in your case....Then you want the grid stopper resistor in series close to the tube socket....
The tough part of this is choosing the correct grid stopper value...
Here is what the trade-off is.... The grid stop resistor makes a Low-Pass filter working into the Miller capacitance of your input valve.. Assuming your using a 12AX7A valve your typicall stage gain of roughly 60 , then you effective Miller Capacitance will be about 152pF......
Now keep in mind that the Pole frequency you choose will be at -3dB and a 45 degress phase shift .....
You want to place the Pole high enough that you don't intoduce any objectionable phase shift in the upper audio band....this you WILL hear and will make the Hi-End system sound cheap...
You don't want to place the POLE to high, then you will not filter out the RF...
So for example if you use a 2.2K grid stopper, your POLE will be at 478kHz and you will have a 1.2 degree phase shift at 10kHz and 2.4 degree phase shift at 20kHz..... this is negligible...
Your 108MHz nuisance of a radio station will have -47dB of attenuation, hopefully this will keep it from being "detected" in your 1st stage of your Phono stage....
You can alway work the value of the grid stopper up till you solve the problem, hopefull it will not alter the sound of your phono stage to were you can hear it...
Chris
Normally the RIAA filter would filter the RF, but in your case the RF is entering the grid and being "detected"...demodulated to audio band...then your phono stage is amplifying the unwanted audio signal....
I also suggest using Hi-End resistors in that location.....first you want to hit the loading resistor of say 47K or 56K in your case....Then you want the grid stopper resistor in series close to the tube socket....
The tough part of this is choosing the correct grid stopper value...
Here is what the trade-off is.... The grid stop resistor makes a Low-Pass filter working into the Miller capacitance of your input valve.. Assuming your using a 12AX7A valve your typicall stage gain of roughly 60 , then you effective Miller Capacitance will be about 152pF......
Now keep in mind that the Pole frequency you choose will be at -3dB and a 45 degress phase shift .....
You want to place the Pole high enough that you don't intoduce any objectionable phase shift in the upper audio band....this you WILL hear and will make the Hi-End system sound cheap...
You don't want to place the POLE to high, then you will not filter out the RF...
So for example if you use a 2.2K grid stopper, your POLE will be at 478kHz and you will have a 1.2 degree phase shift at 10kHz and 2.4 degree phase shift at 20kHz..... this is negligible...
Your 108MHz nuisance of a radio station will have -47dB of attenuation, hopefully this will keep it from being "detected" in your 1st stage of your Phono stage....
You can alway work the value of the grid stopper up till you solve the problem, hopefull it will not alter the sound of your phono stage to were you can hear it...
Chris
Grado recommends a cartridge load of 47K for all of their induced magnet (MI) cartridges including yours - so I would change this as well as add a grid stopper.
In phono pre-amps I have found 220 ohms - 1K to be more than adequate in all cases so far.. And Don't use a grid stopper larger than necessary to do this job because it is another noise source in series with the cartridge. The internal generator impedance is typically something over 600 ohms in series with the coil inductance - so at very low frequencies the grid stopper may generate appreciably more noise than the cartridge if the resistance is much more than the cartridge resistance. (Note the 47K load resistor appears in || with the cartridge and consequently doesn't contribute much LF noise.)
The grid stopper to be effective should be mounted as close as possible to the socket pin.
Small plate stopper resistors (like 10 - 22 ohms) can be helpful, but usually not required with the 12AX7A. Again mount close to the socket.
Take a careful look at your input grounding scheme, if you have star grounded the inputs add a small ceramic cap 0.01uF max from the RCA jack ground connection right to the chassis (one on each channel) - your internal ground wiring has appreciable inductance and isn't a good ground at 100MHz.
The phono stage chassis should be metal, if wood shield the inside with copper tape and connect to the star grounding point. Internal signal wiring should be fully enclosed. (Chassis box with bottom.)
Tube shields can help a lot with tubes having high RP like the 12AX7A.
Bypass filament wiring to ground at the socket with 0.01uF ceramic caps.
And yeah I have designed and built a lot of phono stages over the years. (See my site for some early designs.)
In phono pre-amps I have found 220 ohms - 1K to be more than adequate in all cases so far.. And Don't use a grid stopper larger than necessary to do this job because it is another noise source in series with the cartridge. The internal generator impedance is typically something over 600 ohms in series with the coil inductance - so at very low frequencies the grid stopper may generate appreciably more noise than the cartridge if the resistance is much more than the cartridge resistance. (Note the 47K load resistor appears in || with the cartridge and consequently doesn't contribute much LF noise.)
The grid stopper to be effective should be mounted as close as possible to the socket pin.
Small plate stopper resistors (like 10 - 22 ohms) can be helpful, but usually not required with the 12AX7A. Again mount close to the socket.
Take a careful look at your input grounding scheme, if you have star grounded the inputs add a small ceramic cap 0.01uF max from the RCA jack ground connection right to the chassis (one on each channel) - your internal ground wiring has appreciable inductance and isn't a good ground at 100MHz.
The phono stage chassis should be metal, if wood shield the inside with copper tape and connect to the star grounding point. Internal signal wiring should be fully enclosed. (Chassis box with bottom.)
Tube shields can help a lot with tubes having high RP like the 12AX7A.
Bypass filament wiring to ground at the socket with 0.01uF ceramic caps.
And yeah I have designed and built a lot of phono stages over the years. (See my site for some early designs.)
A common fix to using a smaller grid resistor is to put additional Capacitance at the input of the tube...
For example, a popular Hi-End pre-amp uses a grid stopper of 365 ohms and also puts a 47pF from grid to cathode.....
This is like -34dB attenuation at 108MHz ......
As the last poster mentions... NOISE is a issue ....
But also be aware that the signal current from the cartridge does not pass through the grid resistor ...it passes through the load resistor, 47K ohms.....
Chris
For example, a popular Hi-End pre-amp uses a grid stopper of 365 ohms and also puts a 47pF from grid to cathode.....
This is like -34dB attenuation at 108MHz ......
As the last poster mentions... NOISE is a issue ....
But also be aware that the signal current from the cartridge does not pass through the grid resistor ...it passes through the load resistor, 47K ohms.....
Chris
Kevin is correct about the low gm 'X7 not needing much resistance in the stopper position. 100 Ohm Carbon composition parts, with their bodies right at the socket lugs WILL get the job done. Noise is not an issue with control grid stoppers, as the current present is vanishingly small. Being both non-inductive and non-metallic makes Carbon comp. resistors SUPERIOR in the control grid stopper role.
Take additional steps against RFI. Use shields around the 'X7 bottles. Also, use clamp on ferrites on the cables right next to the I/P RCA females.
BTW, changing the 22 KOhm resistor in the EQ network to 24 KOhms improves RIAA accuracy.
Take additional steps against RFI. Use shields around the 'X7 bottles. Also, use clamp on ferrites on the cables right next to the I/P RCA females.
BTW, changing the 22 KOhm resistor in the EQ network to 24 KOhms improves RIAA accuracy.
With that passive RCA pre-amp...
Yes, changing that 22K to a 24K improves the RIAA acuracy...
As the circuit stands with the 22K resistor you get a 67.5uS Time Constant....
Using 24K moves this time constant to 73uS....
Technically it should be 75uS ......
Also the 3180uS time constant is actually at 3126uS as the circuit stands...changing over to 24K moves this to 3146uS .....
The part that I don't get is where the 318uS Time Constant comes from????
The second stage with it's 180pF produces a 6.27uS and this moves around a wee bit depending on output loading.... I figure if you drop the 180pF to 100pF you will get closer in the ball park like about 3.3uS ....
I am "assuming" that RCA "bean-counters" intended this circuit for low budget mass produced units and were selecting these resistor values based on price and standard chart values, also that it is "close enough" for consumer applications...
I see that the 1KHz mid Band is at +10dB Gain from the first stage...
The second stage contributes +35dB Gain low to mid band...
This makes about +45dB Gain at 1kHz for the whole bloody thing.....unless I goofed somewere in the math
Totally agree on the " Johnson Noise" ..... I don't usually include this in the phono-pre analysis, since the valves thermal noise swamps this out... I do however consider that when designing LNA's in RF applications with noise floors about -137dBm....
Chris
Yes, changing that 22K to a 24K improves the RIAA acuracy...
As the circuit stands with the 22K resistor you get a 67.5uS Time Constant....
Using 24K moves this time constant to 73uS....
Technically it should be 75uS ......
Also the 3180uS time constant is actually at 3126uS as the circuit stands...changing over to 24K moves this to 3146uS .....
The part that I don't get is where the 318uS Time Constant comes from????
The second stage with it's 180pF produces a 6.27uS and this moves around a wee bit depending on output loading.... I figure if you drop the 180pF to 100pF you will get closer in the ball park like about 3.3uS ....
I am "assuming" that RCA "bean-counters" intended this circuit for low budget mass produced units and were selecting these resistor values based on price and standard chart values, also that it is "close enough" for consumer applications...
I see that the 1KHz mid Band is at +10dB Gain from the first stage...
The second stage contributes +35dB Gain low to mid band...
This makes about +45dB Gain at 1kHz for the whole bloody thing.....unless I goofed somewere in the math
Totally agree on the " Johnson Noise" ..... I don't usually include this in the phono-pre analysis, since the valves thermal noise swamps this out... I do however consider that when designing LNA's in RF applications with noise floors about -137dBm....
Chris
SY said:
Ouch... Take a look at the work by Stanley Lipshitz for clue about this issue or take a look at my sloppy, but reasonably accurate RIAA networks based on his work. (Come to think of it that zero might not be there either, if it's the one I am thinking of it is often omitted. SY correct me.. LOL)
An inverse RIAA network you can tweak response against is pretty useful.. I'd get me one of those. There is an old AA article on building on of these. Bear in mind source and terminating impedances for these networks can be critical to accuracy. (Typically 600 ohms source and termination.)
Kevin
First, if your RF problem occurs with the inputs shorted then it is probably within your RIAA circuit.
Years ago I talked with someone at Grado on the phone about a hum problem with one of their budget carts. I learned that they did not believe in shielding their carts. Maybe things have changed at Grado.
I ended up buying an 8MZ Signature Series cart. It has no hum problem.
There may be something about the Grado cart in conjunction with your preamp that is creating an RF antenna.
Mark
Years ago I talked with someone at Grado on the phone about a hum problem with one of their budget carts. I learned that they did not believe in shielding their carts. Maybe things have changed at Grado.
I ended up buying an 8MZ Signature Series cart. It has no hum problem.
There may be something about the Grado cart in conjunction with your preamp that is creating an RF antenna.
Mark
Thank you all so much!
wow this is some amazing advice. I will make that sub in the RIAA section and change the load to 47k and try the various grid-stopped recommendations.
The problem is only present when a cable is inserted in the input jacks... so no I do not think it is in the RIAA section.
This is some great stuff to get started on, thank you all so much. c/
wow this is some amazing advice. I will make that sub in the RIAA section and change the load to 47k and try the various grid-stopped recommendations.
The problem is only present when a cable is inserted in the input jacks... so no I do not think it is in the RIAA section.
This is some great stuff to get started on, thank you all so much. c/
Glad to be of help, remember also that the internal ground wiring is not a ground at 100MHz and you really want to use some caps from the input rca jack grounds directly to the chassis - use small ceramic caps with leads as short as you can reasonably make them. This alone has saved me more than once.
I have tried beads, but found them audible, and IIRC I think they did have a measurable effect on the thd above a couple of kHz. Anyway I usually remove them and replace with small resistors or air core (diy) inductors.
I have tried beads, but found them audible, and IIRC I think they did have a measurable effect on the thd above a couple of kHz. Anyway I usually remove them and replace with small resistors or air core (diy) inductors.
UPDATE: RCA phono pre RF problem
Hi all
so - i implemented a number of the suggestions that the group proposed:
*Changed input load resistor to 47k (was 56k)
*added 820 ohm 1/2w cc resistor between input jack and 1st screen (body of resistor located at socket).
*added 75v .01 ceramic cap b/w heater pins and ground
The level of hum (which was never bad at all) did not change. the snubber caps did not seem to have any effect.
The RF (108 FM) dropped dramatically, but it is still audible... more than i think it should be. it is only present when cables are inserted.
I am using switchcraft-style steel non-insulated input jacks (IE the body of the jack has continuity with the chassis). I am using a ground buss (8 ga solid copper wire) which meets the chassis at the input jacks. Because of this, i did not try the suggestion to connect inout jack ground to chassis thru a .01uf cap... cos the jacks are essentially equivalent to ground in this case, no?
So... any other suggestions? my only other move left is to bump the input grid stopper value up to 2200 ohms.
c/
Hi all
so - i implemented a number of the suggestions that the group proposed:
*Changed input load resistor to 47k (was 56k)
*added 820 ohm 1/2w cc resistor between input jack and 1st screen (body of resistor located at socket).
*added 75v .01 ceramic cap b/w heater pins and ground
The level of hum (which was never bad at all) did not change. the snubber caps did not seem to have any effect.
The RF (108 FM) dropped dramatically, but it is still audible... more than i think it should be. it is only present when cables are inserted.
I am using switchcraft-style steel non-insulated input jacks (IE the body of the jack has continuity with the chassis). I am using a ground buss (8 ga solid copper wire) which meets the chassis at the input jacks. Because of this, i did not try the suggestion to connect inout jack ground to chassis thru a .01uf cap... cos the jacks are essentially equivalent to ground in this case, no?
So... any other suggestions? my only other move left is to bump the input grid stopper value up to 2200 ohms.
c/
YUCK!! Put Gold plated jacks with insulating washers in. Add the 10 nF. ceramic capacitors. Use shielded "twinax" wire between the I/P jacks and the signal circuitry. The shields should be grounded only at the same point the 10 nF. caps. are grounded. The shields do not carry signal. Connect the signal ground bus up near the tube sockets, not at the I/P jacks.
Finally, don't forget the RFI suppressing ferrites on the cables from the TT as close to the preamp as possible. The cables are acting as antennas.
Eli Duttman said:
YUCK!! Put Gold plated jacks with insulating washers in. Add the 10 nF. ceramic capacitors. Use shielded "twinax" wire between the I/P jacks and the signal circuitry. The shields should be grounded only at the same point the 10 nF. caps. are grounded. The shields do not carry signal. Connect the signal ground bus up near the tube sockets, not at the I/P jacks.
Finally, don't forget the RFI suppressing ferrites on the cables from the TT as close to the preamp as possible. The cables are acting as antennas.
Eli, I couldn't have said it better myself!
What he says man...
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.