Right done a simulation of this circuit on LT spice. First the last valve should be 12AU7 or 12AT7 not 12AX7.
Opening up the feedback loop (C107) shows the circuit to be marginally stable at 1MHz-10MHz. So you have probably built an oscillator. The chirp you here is the circuit beating with one of the AM radio stations or the other channel and demodulating the result. Bascially as you bring the gain down to 1x with the RIAA eq it becomes unstable. You need to add a dominate pole to your design. Try a 100p across R108. Better still make R108 and a series cap and r part of the RIAA equalisation, that way the feedback loop does not need to reach unity.
I would add a diode across the grid-cathode of the final valve - and consider running the first and maybe second valve on DC heater.
Oh and I would add 2k2 in series with the output to make sure the feedback path is not influenced by the load of the pre-amp.
Opening up the feedback loop (C107) shows the circuit to be marginally stable at 1MHz-10MHz. So you have probably built an oscillator. The chirp you here is the circuit beating with one of the AM radio stations or the other channel and demodulating the result. Bascially as you bring the gain down to 1x with the RIAA eq it becomes unstable. You need to add a dominate pole to your design. Try a 100p across R108. Better still make R108 and a series cap and r part of the RIAA equalisation, that way the feedback loop does not need to reach unity.
I would add a diode across the grid-cathode of the final valve - and consider running the first and maybe second valve on DC heater.
Oh and I would add 2k2 in series with the output to make sure the feedback path is not influenced by the load of the pre-amp.
Last edited:
I'm also working on a phono amp. Noise and hum was ok but during experimentation I managed to reduce hum to a very low level by placing a shield (A BMX pedal shaped like a heavy tube) over the first valve. This has reduced hum to a near non existent level. (unable to see hum on scope at 2mv setting)
I was surprised by the level of hum pickup with the valve un-shielded. (even with the mains transformer in its own shielded box.) Just a thought
No, the final phono amp will not use the BMX pedal as a shield
I was surprised by the level of hum pickup with the valve un-shielded. (even with the mains transformer in its own shielded box.) Just a thought
No, the final phono amp will not use the BMX pedal as a shield
Yep capacitance on the output to ground will add a second pole and take it into instability. 2k2 in series with the output lead only (not the FB path) should stop this. A cap across the second plate resistor could be used to form a dominate pole so the system is unity gain stable. If its oscillating the hum also may change in very strange ways.
Hi Everybody,Some much to respond to! You are all being very generous!
First of all, I really think I need to rework the layout. The way I have it now is very cramped. (see photo)
1) Analog_sa:
Since you have had this circuit working, how did you have your filaments set up?
Right now all 3 Filaments are parallel connected DC elevated 30V.
2) baudouin0:
-Thanks for sim-ing the circuit! I will try 2k2 in series with the output. Makes sense. I will also, try a cap across 108.
-The cathode to grid diode was also suggested by petrtub. So the orientation would be Diode cathode to tube cathode and diode anode to tube grid?
-I have tried both 12au7 and 12ax7 in the cathode follower position. Not much changes other than the voltages rose slightly with the 12au7 in place.
-Yes the hum was not linear relative to volume. It seemed that the final stage was having a strange interaction with the volume pot in the preamp that follows. At low volume the hum was almost acceptable, and one could hear the signal (test record) clearly. As the volume was raised the ratio of signal to noise changed drastically. By about 10 o'clock (right where one might do most listening) all that could be heard was hum. no signal at all.
3) Paul Joppa:
-The -23V at V2 plate was a typo. I meant to type grid. Sorry. But why did I get a reading of -23V at V2 plate relative to the signal ground?
That one stuck me as odd.
-The current readings were taken with a DMM measuing current across the plate and cathode resistors. I probably did this in correctly!
Yes. When the grid is made positive (and the cathode still is cold and non-conducting) the grid will be many volts positiv vs. the cathode. This is where the diode should conduct. The current will maximally be (B+ / Rcathode). When the cathode has been warmed it will conduct and the cathode
will raise slightly above the grid, the diode will block and be by all
means 'out-of-circuit'. Smallest diode ( 1n4148) will do and have low
capacitance in block state.
When using 12AU7 better reduce the cathode resistor to , lets say, somewhere
around 50k. This will also reduce the output impedance.
There is still the issue with filament <> cathode voltage , if no separate
source can be found it could be better to lower the grid voltage by increasing
the plate resistor of the previous stage OR make the cathode follower AC
connected. 100Volt filament <> cathode is what would aim for.
I think you are allowed up to 250V cathode - filament on a 12au7. the 12ax7 is lower. First see if you have oscillation and stop this.
-Yes the hum was not linear relative to volume. It seemed that the final stage was having a strange interaction with the volume pot in the preamp that follows. At low volume the hum was almost acceptable, and one could hear the signal (test record) clearly. As the volume was raised the ratio of signal to noise changed drastically. By about 10 o'clock (right where one might do most listening) all that could be heard was hum. no signal at all.
Oscillation.
As far as layout is concerned you may be better with a combination of tag board and point to point as the number of passives is quite high.
-Yes the hum was not linear relative to volume. It seemed that the final stage was having a strange interaction with the volume pot in the preamp that follows. At low volume the hum was almost acceptable, and one could hear the signal (test record) clearly. As the volume was raised the ratio of signal to noise changed drastically. By about 10 o'clock (right where one might do most listening) all that could be heard was hum. no signal at all.
Oscillation.
As far as layout is concerned you may be better with a combination of tag board and point to point as the number of passives is quite high.
Last edited:
Sorry, but per the voltages given in the schematics the CF output stage is completely off. I'm well astonished that you hear anything at all. Due to it's much higher transconductance, I'd prefer a 12AT7 over the 12AX7 here, and also over the 12AU7 that has been suggested before.
Best regards!
Best regards!
Analog sa: Do you recall at all how your heaters were arranged when you built your 834p-like phono stage?
Thanks, again, to everybody for all your help!!
Interesting.
So, adding the 2k2 resistors in series with the output helped things a lot.
Hum and buzz are tamed and I managed to listen to a whole side of an LP.
Adding the 100pf caps in parallel with the V2 plate resistors made the unit into a very pure clear single frequency sine wave generator. Something like 1200hz-ish. Nice pure tone, though. Maybe this indicates other problems.
Diodes added g->k on the cathode follower.
Still getting the loud chirps when the preamp output is temporarily disconnected (i.e. turning the source selector on the following preamp) or if the tonearm ground lead is connected or disconnected.
At this point I am going to tear the phono amp down and start again. (Petertub: I promise to also address the filament issue.)
This may take a while. Hitting a busy period for the next couple months.
Thanks so much to everyone who chimed in!
I will post the solution when I find it sometime in the future. I hate threads that end without resolution.
So, adding the 2k2 resistors in series with the output helped things a lot.
Hum and buzz are tamed and I managed to listen to a whole side of an LP.
Adding the 100pf caps in parallel with the V2 plate resistors made the unit into a very pure clear single frequency sine wave generator. Something like 1200hz-ish. Nice pure tone, though. Maybe this indicates other problems.
Diodes added g->k on the cathode follower.
Still getting the loud chirps when the preamp output is temporarily disconnected (i.e. turning the source selector on the following preamp) or if the tonearm ground lead is connected or disconnected.
At this point I am going to tear the phono amp down and start again. (Petertub: I promise to also address the filament issue.)
This may take a while. Hitting a busy period for the next couple months.
Thanks so much to everyone who chimed in!
I will post the solution when I find it sometime in the future. I hate threads that end without resolution.
You have built an oscillator. Sorry the 100p on LT spice does make it more unstable. However the original design with output resistor should work just fine. If you do redo layout keep the input away from the output and use screened cable for both.
This makes a good read:
miscellaneous-schematics1
Sorry we could not solve it.
This makes a good read:
miscellaneous-schematics1
Sorry we could not solve it.
Hey! I haven’t given up! I’m just not going to be able to work on it for a bit.
Everyone’s advice has been extremely helpful! Thank you!
Besides, I’m firmly convinced it was my own bad layout that caused the problems.
I know a lot more now than when I started, so that is an added benefit.
When rebuilding, be sure to install the the "output resistor" directly at the follower's cathode pin. Cathode followers make excellent oscillators, with their heavy feedback, and really benefit from cathode "stopper" resistors. Lack of 'em has even given c. followers a bad name.
BTW, I really love your elegant anode follower circuit.
All good fortune,
Chris
BTW, I really love your elegant anode follower circuit.
All good fortune,
Chris
Not sure why you ask, this is unrelated to your oscillating issue. I always use regulation on the heaters but cannot remember if it was voltage or current and whether each tube had its own. Likewise the dc bias.
In fact most of the time i spent with this circuit was dedicated to air dielectric and diy teflon caps for the riaa network. It was interesting to say the least. Btw i quickly gave up on the ECC83 CF and used a 5687 instead.
Your main issue is the cramped build which creates too much parasitics. Otoh, this is not a high nfb circuit, on the contrary. How is your power supply? Regulated? Input stage decoupled well?
Before rebuilding you should try to narrow down the source of oscillations. Per example you can remove entirely the input tube and make sure the rest is stable.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.