The LEDs are deep within a chassis with ventilation holes on the sides. Little likelihood of room LED light penetrating but I could turn off all the room lights and see if that effected things.
Yes. The input stage is the most important of course. Bypassing the LED will buy you a few extra dBs of SNR, assuming everything else is blameless.
The grid resistor of 300r is doing almost nothing with low dynamic capacitance of the 1st stage tube...
https://www.diyaudio.com/community/threads/input-capacitance-of-a-triode-connected-d3a.375614/
and
http://www.valvewizard.co.uk/gridstopper.html
.
Can be calculated more precise with given inter-electrode capacitances
Cdyn= (I-AI+1)xCak + Cgk
it is roughly 70-80pF
Rgrid=1/(2 pi x Fo x Cdyn), where Fo is -3db cuttoff frequency.
for 100KHz Fo Rgrid is about 21Kohm for 200KHz is about 10Kohm and so...
with these 300ohms Fo=530MHz 🙁 not a much of stoping oscillations?
.
Grid stopper at the 2nd tube, (also 300ohms) is unnecessary because in-front of the tube HF filter already exists and grid stopper is in the same time firs R in network...
https://www.diyaudio.com/community/threads/input-capacitance-of-a-triode-connected-d3a.375614/
and
http://www.valvewizard.co.uk/gridstopper.html
.
Can be calculated more precise with given inter-electrode capacitances
Cdyn= (I-AI+1)xCak + Cgk
it is roughly 70-80pF
Rgrid=1/(2 pi x Fo x Cdyn), where Fo is -3db cuttoff frequency.
for 100KHz Fo Rgrid is about 21Kohm for 200KHz is about 10Kohm and so...
with these 300ohms Fo=530MHz 🙁 not a much of stoping oscillations?
.
Grid stopper at the 2nd tube, (also 300ohms) is unnecessary because in-front of the tube HF filter already exists and grid stopper is in the same time firs R in network...
Last edited:
Another thing I spot
is that D3a have internal resistance in triode mode of 1.9K. In parallel with Rload it is smaller value of resistance as the generator.
Passive riaa net is about 22K of impedance as load. This is too low riaa net value. And inserting a signficant limiter/compression effect.
Simply loading too hard the generator...
Should be say as 200K. 100X higher. I suggest to make 10X bigger values of Rs and 10x smaller of Cs in the net. Or recalculate net with 200-220K First resistor.
is that D3a have internal resistance in triode mode of 1.9K. In parallel with Rload it is smaller value of resistance as the generator.
Passive riaa net is about 22K of impedance as load. This is too low riaa net value. And inserting a signficant limiter/compression effect.
Simply loading too hard the generator...
Should be say as 200K. 100X higher. I suggest to make 10X bigger values of Rs and 10x smaller of Cs in the net. Or recalculate net with 200-220K First resistor.
Whoa - that is beyond the scope of this course for me. I just want to eliminate the his in one channel - not modify the whole thing unless that's what I need to do. I do thank-you for your interest though.
Aren't you mixing up oscillations and RF filtering?
Yes.
But I think that the circuit has other issues.
First try to replace diodes in the catode with RC. Like @Merlinb said few posts before.
OR You cam use 1.5V battery later after the check.
.
I think that Rload of D3a of 47K is too big value. 6.8K to 10K is more appropriate.
.
In this arrangement, position of passive riaa network, all the elements in net are exposed to high voltage.
There is no need for this.
Put the 1uF Co (high voltage) imeadetley after first tube at the anode. And one high value R of 1meg - 2.2meg to ground after coupling C.
Then put ria netwotk.
No nedd for grid stopper or input R fir 2nd stage, because tube grid is conected to ground via this 1-2.2meg.
.
With 300ohms there are no any rational f for filtering. So there are open door for oscilations too
For oscilations preventing maybe it is better to use some small indictance with low Rdc. But in this case there are no RF filtering...
.
Anyway it should be mesured if it is possible...
It is not a hardware or any big thing modification. But only few elements and arrangements inside the existing classes. 20min job...
Start with present state, measure anode voltages, bias voltages, calculate or measure current and check with anode curves with measured values where are operating points? And is there some difference in channels.
D3A anode load resistor is 4K7, not 47K ( reading the original schematic when published at other forum ).
For testing just use the resistor. The the plate load it decent enough the cap by-pass won't matter so much.
The main purpose is to rule out the LED's as a source of hiss. If you don't get hiss, then we know its most likely one of those LED's
If you still get hiss, then its got to be somewhere else. Hiss is a bugger. I much prefer to figure out where hum is coming from. or Johnson noise for that matter. That stuff is much easier to de-bug than hiss.
You did swap the tubes, right?
The main purpose is to rule out the LED's as a source of hiss. If you don't get hiss, then we know its most likely one of those LED's
If you still get hiss, then its got to be somewhere else. Hiss is a bugger. I much prefer to figure out where hum is coming from. or Johnson noise for that matter. That stuff is much easier to de-bug than hiss.
You did swap the tubes, right?
Looking at Zoran's explanations, he might be onto something. but it won't cause hiss.
Trust me on this one. its an easy thing to swap out. instead of LED's use a resistor - what is the current on your D3A ? 15ma or something? What is your bias point? 1.5 V? Use ohm's law. what is that HLP6000 ? is it a red LED ?
EDIT: AHA. after MUCH searching I find a 1.6V RED mini LED called an HLMP-6000
https://www.mouser.ch/ProductDetail/Broadcom-Avago/HLMP-6000
Wow, these are expensive little buggers. Why build expensive when you can do it way cheaper? (Mrs. soulmerchant asks me this question with every build)
Ok so swap out that little red bugger with a 100 OHM metal film or carbon resistor. That should be good enough for rock and roll. If the hiss doesn't go away then you need to by-pass the stack of the next stage (where I suspect the culprit may be).
Actually, WAIT.......... why the heck do you need to bias the next stage at 4x 1.6V??? Ah I see. Its the 5687 and you should have something like 150V on the plate ? right? with a bias of 6.4V - OK... THEN please MEASURE That plate voltage on the 2nd stage and tell us what that is. Do it for EACH channel.
BTW - next time pls show us the COMPLETE schematic with voltages. This will save a LOT of speculation as well. If you are at a crazy bias point and your plate voltage is way too high then THAT will be the reason for Hiss. You might have fried your CCS on the second stage. THAT would make sense and definitely cause hiss.
Ian
Trust me on this one. its an easy thing to swap out. instead of LED's use a resistor - what is the current on your D3A ? 15ma or something? What is your bias point? 1.5 V? Use ohm's law. what is that HLP6000 ? is it a red LED ?
EDIT: AHA. after MUCH searching I find a 1.6V RED mini LED called an HLMP-6000
https://www.mouser.ch/ProductDetail/Broadcom-Avago/HLMP-6000
Wow, these are expensive little buggers. Why build expensive when you can do it way cheaper? (Mrs. soulmerchant asks me this question with every build)
Ok so swap out that little red bugger with a 100 OHM metal film or carbon resistor. That should be good enough for rock and roll. If the hiss doesn't go away then you need to by-pass the stack of the next stage (where I suspect the culprit may be).
Actually, WAIT.......... why the heck do you need to bias the next stage at 4x 1.6V??? Ah I see. Its the 5687 and you should have something like 150V on the plate ? right? with a bias of 6.4V - OK... THEN please MEASURE That plate voltage on the 2nd stage and tell us what that is. Do it for EACH channel.
BTW - next time pls show us the COMPLETE schematic with voltages. This will save a LOT of speculation as well. If you are at a crazy bias point and your plate voltage is way too high then THAT will be the reason for Hiss. You might have fried your CCS on the second stage. THAT would make sense and definitely cause hiss.
Ian
Last edited:
Seems to work just fine in SY's "His Master's Noise" -- mine hasn't failed yet. Plate dissipation of the D3a is 4.5W
https://www.diyaudio.com/community/...a-thoroughly-modern-tube-phono-preamp.163570/
Increasing the build-out resistor to 220k will cause havoc with noise from the first stage.
Jack
Do you run the phono stage with the enclosure buttoned up or open as in the photos you posted?
Bottlehead uses that very LED on several of its kits. It would surprise me if they ever caused his, but it may be worth checking over in their forums. The Crack comes with those LEDs.
I would hesitate to change any operating conditions of the phono stage before locating the source of the hiss. Start with the simplest thing to rule out and work up.
My experiences with oscillation have sometimes been delayed--taking a while to manifest after turning on. For me, though, oscillation has always sounded high-pitched, crackly, and strident.
If your CCS is "fried" then you won't have sound on that channel.
I would hesitate to change any operating conditions of the phono stage before locating the source of the hiss. Start with the simplest thing to rule out and work up.
My experiences with oscillation have sometimes been delayed--taking a while to manifest after turning on. For me, though, oscillation has always sounded high-pitched, crackly, and strident.
If your CCS is "fried" then you won't have sound on that channel.