The pop happens when I switch to the tubed phono stage. The cap is about
5uF and there is 1M resistor (from memory) from the phono output to ground.
The tube side of the cap has about 100V DC on it, the output side is always
pulled to ground with the 1M resistor. So the cap should not have to recharge.
One possibility is the LC circuit created by the phono output cap and the inductance
of the volume autotransformer.
I will try to put 10kOhm on phono output and see if the problem is film cap leakage.
Simon,
Then how about two film caps in series? Wouldn't that reduce any leakage to an infinitesimal level?
Then how about two film caps in series? Wouldn't that reduce any leakage to an infinitesimal level?
The circuit works, but with my sloppy layout with large loop areas and bulky battery banks (three six-packs of AA cells) it's picking up too much local interference noise. So I will have to find a box to get everything into and enclosed, thus it may be a while (I've managed to use up a bunch of old chassis so may actually have to purchase some). However, one benefit noted already is higher gain (increase of about 0.78dB) replacing the pullup-pulldown R's with the floating I source, which accords fairly well with simulation results. This is due to the input device drain currents flowing into the folded cascode parts nearly completely, rather than shared a bit with the previous resistors.
I think I joked in here somewhere that some Audio Note copper foil/paper/oil caps, prized by an associate for good sound, had so much leakage that they were the first ones used as coupling caps that needed a d.c. servo.
Someone mentioned thereafter that AN has migrated back to film caps in recent products.
If modeled as a parallel resistance, and if the caps are identical, then two would just halve the leakage current. However if a pulldown be placed between them, that would help a lot.
If one has a sufficiently high impedance voltmeter, measure the self-discharge time of an isolated cap using brief sampling of the voltage at intervals, after charging it to, say, 100V. A lot of DMMs are 10Mohm, which means the sampling has to be short.
As I recall a typical spec for a good film cap is discharging to 1/e of the initial voltage in 10^4 to 10^8 seconds. That latter is over 1100 days.
Leakage currents rarely obey Ohm's law: sometimes they do, more or less, sometimes they just behave like currents, ie putting two in series changes nothing, but they can also behave like VDR's, meaning that placing two in series will substantially reduce their level.
Things can be more complicated than that, and there can be a mix of the above-mentioned behaviors, especially if the two currents are not absolutely identical
Audio Note Japan used to make silver foil - Mylar caps by hand and use
them in their products. I believe AN UK now uses similar caps.
one confirmation
One conjecture has been proven, despite the poor layout at the moment: one can lose the bypassing caps for the folded cascode parts using the floating I source, without a discernible change in the noise.
Vac,
Current leakage in capacitors is very non-ohmic. So two in series with a shunt resistor will work, two by themselves will not. Found this out the first time I built a bench top power supply, the final mylar film capacitor at the output terminals showed a draw of 10 ua on the current meter!
Brad,
A secret source of isolation metal boxes are the tin cans from Walgreens that the $1.99 (on sale) Danish butter cookies come in. Around a foot in diameter. Very handy source of cheap rugged bins.
The other secret is Hershey's syrup in squeeze bottles cost less than purpose sold wood glue bottles.
ES
Current leakage in capacitors is very non-ohmic. So two in series with a shunt resistor will work, two by themselves will not. Found this out the first time I built a bench top power supply, the final mylar film capacitor at the output terminals showed a draw of 10 ua on the current meter!
Brad,
A secret source of isolation metal boxes are the tin cans from Walgreens that the $1.99 (on sale) Danish butter cookies come in. Around a foot in diameter. Very handy source of cheap rugged bins.
The other secret is Hershey's syrup in squeeze bottles cost less than purpose sold wood glue bottles.
ES
But what would I do with the cookies? I reserve my empty calorie allotment for wine.
Bcarso,
Is there a particular reason you chose NiMH over Li batteries or was it just that you had them at hand? Any advantage of one over the other? Are you doing this just to eliminate the power supply noise for your testing or are you going to use batteries in your final design? I suspect it is only for evaluation purposes but wanted to ask.
Is there a particular reason you chose NiMH over Li batteries or was it just that you had them at hand? Any advantage of one over the other? Are you doing this just to eliminate the power supply noise for your testing or are you going to use batteries in your final design? I suspect it is only for evaluation purposes but wanted to ask.
If you give them to the neighborhood children you just might get arrested. Feed them to the birds and a woman with lots of cats might say hello. So winos are the safest bet.
I am interested in just how you are going to make a really quiet current source with the many ma needed for this circuit.
Just dump the cookies in the trash or put them in a bag and give them to a homeless person, they can use the empty calories!
- Status
- Not open for further replies.