| burnedfingers |
I purchased what is supposed to be a clone of a Jadis line stage that contains 3) 12AX7's per channel. It took about 15 min to stuff
as it was laid out pretty well.
It sounds good except I can't get rid of the hum problem. I am running a R*C*R*C power supply with 360volts going to feed the board. I am running 6.3VDC with a 27,000mfd cap for filtering to feed the heaters.
So far I have listened to one channel with the power and heaters unterminated on the other side.
The board has all the grounds tied together for each channel so I would assume this wouldn't be the problem.
Its actually 2 separate channels on one board.
Any ideas?
I will try to send a schematic of it |
|
|
| SY |
| How are you treating the tied-together board grounds with respect to input socket ground, power supply ground, and chassis ground? |
|
|
| burnedfingers |
SY
So far the board isn't mounted to a chassis. I have soldered RCA jacks on for the input and output terminations. The ground on these is commoned to the ground buss. My power supply ground is soldered to the common ground buss. The negative off my Bridge (heater supply) goes to the common/chassis of the power supply.
I have a stand alone power supply that I am using on it right now before building the finished product.
Do the values look to be in line? Would you suggest any changes to the circuit? |
|
|
| SY |
The values look fine- this is just one more variation of the old Marantz preamp. Check voltages at the plates of the first two tubes and the cathode of the CF to confirm that everything is working as it should.
I'm somewhat confused about the ground bus- is it indeed a bus or are the grounds tied together in a star? If the bus has two ends, so to speak, that may be the source of your hum. |
|
|
| burnedfingers |
I will try to send a picture of it. It might be easier to view a problem if you can see the board. I will heat up the garage and check the voltages in the different sections.
Note. Like I mentioned the board is laid out with two separate channels. It wouldn't take much to score and snap the board for mono sections.
What is your feeling on the feedback loop? |
|
|
| jaudio |
| Where are the B+ supply bypass filter caps I dont see them |
|
|
| SY |
| quote: | Originally posted by jaudio
Where are the B+ supply bypass filter caps I dont see them |
C108/208 and C109/209. Looks like they're mounted on the underside- I see solder joints. |
|
|
| SY |
| From what I can make out in the picture, there are two separate buses, one for each channel. They should only connect at a star ground at the input socket grounds as should the return for the B+ and heater feed. Make sure there are no unintended loops on the outputs, too. |
|
|
| burnedfingers |
Since this is supposed to be a Christmas present for a friend I wanted to make it easier for me and I mounted the caps on the bottom of the board thus making 6 less holes for me to drill in the top plate. I am going to use stand offs to mount the board under the top plate. I figured 6 tubes don't look too bad but add 6 caps to it and it starts to get busy.
Problems...
First I don't know/remember how to calculate the cap in the feedback loop. I question the 1P value. I have I believe 330 pf in there now.
Secondly, there is a voltage pattern on the output that keeps repeating its self. The voltage ramps up to several volts positive then back to zero followed by a voltage ramp the negative way.
What I did right or wrong... I figured I would change the value of the 470K R102 resistor to a smaller value. I used 27K because I had one handy and the voltage ramp stays in the mv range now.
R103 voltage changed from .9V to 1.3V R104 changed from 17.4V
to 20.4V
Oh, I added by pass caps on each tube heater of .01mfd.
Any comments? Ideas? |
|
|
| janneman |
| quote: | Originally posted by SY
C108/208 and C109/209. Looks like they're mounted on the underside- I see solder joints. |
They can't be under - not enough space, unless the wrong value caps are mounted. May be the problem.
Jan Didden |
|
|
| burnedfingers |
Jan
The parts sheet calls for 6) 100mfd/450V caps. I have them below the board. |
|
|
| jaudio |
| I see the bypass caps on the schematics but not on the picture of the board. When bypass cap dry out or fail,you'll get hum |
|
|
| janneman |
| quote: | Originally posted by burnedfingers
Jan
The parts sheet calls for 6) 100mfd/450V caps. I have them below the board. |
OK, I see. I thought they were different ones, the board looked low on the sheet in the pic.
How loud is the hum? Where is your volume control, before the preamp? Did you try with the input shorted?
Jan Didden |
|
|
| burnedfingers |
| Any comments on the feedback change I made? |
|
|
| janneman |
Well, the 330pF with the 100k will roll your freq response off starting at 5 kHz or so. This will defenitely sound dull -;)
In fact, this may cause the ramping, what is the ramp frequency?
The 1pF is not unreasonable for ultrasonic roll off. Try the smallest cap you can find, like 5 pF.
Jan Didden |
|
|
| burnedfingers |
So I must have had a high frequency osolaton with the 470K resistor in the feedback loop?
Am I more in line with the different feedback resistor value?
Can't I leave it off (1P cap in feedback loop)since the tube amp should cover the low frequency roll off? |
|
|
| burnedfingers |
| Volume control is before the preamp. Tried with input shorted and still hum. Well, it(hum) wasn't super loud but you could tell it was there. |
|
|
| SY |
| The 1 pF cap in the feedback loop is for high frequency stability. It's easiest to use a gimmick cap, which is just two one-inch lengths of wire twisted together. Heck, the feedback resistor probably already has a pF across it just in stray... |
|
|
| burnedfingers |
Sy
Would you suggest trying it without the cap? I have never seen a
gimmick cap so I'm not sure on making one.
What are your feelings on the resistor change I made? R102 470K to 27K? |
|
|
| janneman |
If I may (sorry SY), why did you make the change, and why that value?
Jan Didden |
|
|
| burnedfingers |
Jan
I don't have many values on hand but I found that this one got rid of the problem of the voltage ramping up and down on the output. It looked like a strange oscolation(sp). When I disconnected it the problem went away so I assumed it sure wouldn't hurt to try a different value. When I tried the 27K it acted correctly without a ramping problem and still gave me the same output. just running the output to the scope I get 29V rms out with 3 volts in, the same as before.
I made a twisted wire pair from 1" of #22ga wire and twisted the two together and soldered one end of each and floated the other end. Is this correct to replace the cap in the feedback loop?
Sound ...
Its ballsey with a lot of punch. |
|
|
| janneman |
Yeah, that 'cap' seems OK. It was a standard practise in the hf world long time ago, don't know if they still use it.
I thought that the change of R102 to 27k would *increase* the problem, as it give stronger bootstrapping to the input, which is a form of pos feedback. But there probably was some phase shift through that 330pF turning it back into neg feedback. Funny how complex simple circuits can be...
Glad you enjoy the sound - that's the final arbiter, isn't it.
Jan Didden |
|
|
| burnedfingers |
Jan
Without the 330pf cap there I still had the same problem. No matter what I did it kept ramping DC on the output. I'm still not satisfied with what I did. Is there a better solution?:smash: |
|
|
| SY |
| Is that "ramping" an up and down drift of the DC with a period on the order of a second or two? |
|
|
| burnedfingers |
| The ramping process takes more on the order of about 10 seconds. I am wondering if it would be of benefit to change the 470K R110 value to 100K? |
|
|
| fdegrove |
Hi,
| quote: | | I am wondering if it would be of benefit to change the 470K R110 value to 100K? |
It would speed up discharging the cap, yes.
As a side effect you may get somewhat better dynamics as well...
Where the heck does that dinosaur of a circuit come from anyway???
Geeeezzz.........:xeye:
360V on that CF's plate seems rather steep too....
If you'd use a split cap configuration on the output you could at least do away with the input cap (C110).
That should improve the sound.
I think Jadis reworked the thing later on to use a lower B+, a 12AU7 CF as the first triode followed by a 12AX7 direct coupled to a 12AU7 CF.
Still needlessly complicated IMHO. |
|
|
| burnedfingers |
Hi Frank
Can you point me towards a link for the Jadis schematic?
Quote:
Where the heck does that dinosaur of a circuit come from anyway???
Geeeezzz.........
Leave it to me to drag in something completely out of date. |
|
|
| fdegrove |
Hi,
| quote: | | Can you point me towards a link for the Jadis schematic? |
Not sure where it was I found it...
Probably one of those Asian sites I guess...
Not even sure it's an actual Jadis circuit.
It did say so where I had it from but I'm starting to doubt these folks' sanity lately...(I'll leave it up to a guess as to whom I'm refering to......)
I took the liberty to add some comments.
No idea what it sounds like, you'll probably be running it with the handbrake firmly in the up position given the gain of the thing........
Cheers, ;) |
|
|
| burnedfingers |
I changed the R102 value back to 470K. I changed the R110 value from 470K to 100K. I now acts normal and the DC offset stays in the 40-50mv range.
With the grounding at one point on the circuit board I still have hummmm.
Any ideas? |
|
|
| SY |
Hum in both channels?
Hum with inputs disconnected?
Hum that stays the same in level when tubes are swapped around? |
|
|
| burnedfingers |
SY
I have hum in one channel as I haven't fired the other one up yet until all bugs are gone. It will have same hum I am sure.
Hum stays at the same level and makes no difference with tube swap. I have a box of 12AX7 tubes and have tried 9 with same result.
I would think that with DC on heaters it would be quiet.
I need a tall glass of red wine, maybe the whole bottle. Good thing I am on vacation now.
Any ideas? |
|
|
| SY |
Yes, I'd recommend a nice 2001 Michel Ogier "La Rosine" Syrah Vin de Pays.
If you feel ambitious, disconnect the feedback loop, short the input of the cathode follower to ground and see if you still have hum. |
|
|
| fdegrove |
Hi,
| quote: | | I would think that with DC on heaters it would be quiet. |
Not necessarily...
I'd work with both channels live, the B+ is probably common to both channels and with only half the load it may be stressing a tube.
Is the heater supply common to both channels? Maybe it's earthed at the socket side of the channel you're not using?
Cheers,;)
P.S. May be best to finish the bottle after the problem's been solved............:angel: |
|
|
| burnedfingers |
Frank
The channels are totally separate from one another. I could score the board down the center and seperate them. The heaters aren't hooked up on the empty channel nor is the power supply. The tubes are out also.
I will wait on drinking the wine until the line stage works correctly.
SY
It sounds like a fine wine but here in Iowa there isn't too much other than ripple available in the wine section. I will look however.
Here a wine is great if it has a cork in the bottle.
If you feel ambitious, disconnect the feedback loop, short the input of the cathode follower to ground and see if you still have hum.
Ah, you are thinking that with about 20db of gain my problem may lie there? I will try it. |
|
|
| SY |
| No, simpler than that- I'm trying to see in which stage the hum is originating by working backwards from the output. Thinking a little deeper (i.e., clearing my throat after saying something stupid), I realize that you can't actually take the CF grid directly to ground. Best you can do is take the second stage tube's grid to ground. So, OK, let's start there. |
|
|
| burnedfingers |
Ok, second stage's grid to ground. Still have noise.
I changed the feedback resistor value from 100K to 27K and the noise went down some in level. The output is down from 29V with 3 volts in to about 11volts out. It is more of a buzz now. |
|
|
| gingertube |
Check the 12AX7 CF output tube cathode voltage. If its above 100V you may need to cut some tracks etc. to float its heater supply (use the raw AC). By "float" I mean really float it with just a cap from one side to 0V or alternatively tie it to say +80 to +100V DC.
Cheers,
Ian |
|
|
| burnedfingers |
I replaced the 100K feedback resistor. I eliminated the DC on the heaters and tied a 100ohm resistor from each 3.15v leg to gtound. The noise is less now than with the DC on the heaters.
The heaters are wired using 22ga wire and don't have their own traces so there is nothing to cut. I will have to look up the voltages that I have written down for every stage. If my memory is correct there is about 100v cathode voltage.
Are you saying to inject 80-100 VDC into the heater voltage? |
|
|
| fdegrove |
Hi,
Joe,
This is what I suspect is causing the problem:
| quote: | | 360V on that CF's plate seems rather steep too.... |
and is what I think Gingertube is concerned about as well, namely heater to cathode leakage.
Cheers, ;) |
|
|
| burnedfingers |
Hi Frank
Well, I too was concerned with that much voltage so I lowered it for trial purpose. It still had no effect on the hum. I can get rid of the hum if I pull out the CF tube but then what good it the linestage.
Voltages across the resistors: With 360V feed.
R103 .908, R104 17.4, R105 198, R107 .760, R108 246, R109 112
To me it would appear that the 112V was too high across R109, and I don't think the cathode voltages are correct in the first stage. Can I rebias both the cathode follower stage and the first stage or am I barking up the wrong tree? Can I re do the first stage and remove the feedback connection and jumper across R104? That would put R102 to ground(470K) and give me a cathode resistor of 1.6K.
Is Ian wanting me to make a voltage divider and throw B+ into the heater circuit of the cathode follower only or the whole heater supply? |
|
|
|
