ac ripple on screen grids Ultralinear

[Rattler66]

Just to be sure - you can't actually ground the input grids (pin 5) as that would force the bias voltage seen by the output tube to 0V. You could AC short each pin 5, via a cap, to cross check that there was no ripple getting to the input grids. You certainly need to confirm that the bias voltage is not the hum injection point - it often is.

You could temporarily add an RC filter to 'A' supply going to OT CT - a means of checking whether the hum was mainly related to B+ supply.

just to clarify I only ac grounded the grids through a .01u cap to ground, so im not grounding the dc just the ac.

and yes I will double check the bias circuit (I would think ac grounding it also-right?)

mega kudos........doing a temporary rc filter would be a great way to diagnose, I didn't think of that..thanks

[Rattler66]

well I put in a temporary rc filter between the rect and the first node and it did nothing. it was a 60u cap going into 2 100 ohm/10 watt resistors to my existing plate/screen node. so now I'm doubting filtering. I'm getting 3.8vac at the first node then 2.7vac at the second (this is on the cap can) .

again if I ac ground the screens/plate (both coming off the ot primary) I can get the plate/screen voltage to stabilize. if I remove the cap the meter just beeps and jumps around trying to read the voltage. so without that cap there I cannot even determine the proper bias because I can read the current (1ohm resistor at cathode) but not the plate voltage.

I'm starting to wonder if im getting any kind of rf oscillation at the screen and should I up the screen resistor to a 1k?

[trobbins]

A 0.01uF cap bypassing pin 5 to ground is too small to be a significant bypass for 100-120Hz. Perhaps temporarily connect 1uF to each input grid.

Just to clarify - the 200 ohm would connect the power supply to the OT CT, and the new 60uF would connect to the OT CT. You can model this in PSUD2 to check the anticipated levels of ripple at the power supply side of the 200 ohm, and at the OT CT side (which should be much lower ripple).

Remove the screen stoppers from the UL transformer, and use the temporary 200 ohm/60uF as a new screen supply.

Also check the DC resistances of the OT primary windings, starting from a plate, and measuring to its screen, then CT and then the other screen, and other plate.

[loudthud]

There is a slim but real chance that the amp is oscillating at an ultrasonic frequency. That could cause your meter to go crazy. This is where a scope could save a lot of time, but since you don't have one, try setting your meter to AC Volts and connect across the load. The meter can't tell the difference between an oscillation and the hum, and the frequency of oscillation might be so high that the meter won't read it with any accuracy, but see what reading you get with and without the cap on the screens.

Do you have any kind of resistor between the UL tap and the screen? Any grid stopper? These resistor suppress any tendency of tubes to oscillate.

Edit: An AM radio makes a good detector. Place one next to the amp tuned to the low end of the dial. Any difference when the amp is on or off?

[pdf64]

..I'm starting to wonder if im getting any kind of rf oscillation at the screen and should I up the screen resistor to a 1k?

I think that a schematic may be helpful
A good photo of a hand drawn sketch could be fine.

[Rob's Radio-Active]

I think that a schematic may be helpful
A good photo of a hand drawn sketch could be fine.

He posted one on the previous page. I'll put it here again. Click for large view. (Anyone know how to add a full sized inline image?)

[g1]

(Anyone know how to add a full sized inline image?)

See here there: http://music-electronics-forum.com/t41874/#post419628

[Tom Phillips]

Note that if you use any method that relies on an external URL then you image will be gone from the thread if the host ever disappears. The MEF attachment hosting doesn't have a perfect history but, in general, it seems better for longevity than external hosts.

[Rattler66]

There is a slim but real chance that the amp is oscillating at an ultrasonic frequency. That could cause your meter to go crazy. This is where a scope could save a lot of time, but since you don't have one, try setting your meter to AC Volts and connect across the load. The meter can't tell the difference between an oscillation and the hum, and the frequency of oscillation might be so high that the meter won't read it with any accuracy, but see what reading you get with and without the cap on the screens.

Do you have any kind of resistor between the UL tap and the screen? Any grid stopper? These resistor suppress any tendency of tubes to oscillate.

Edit: An AM radio makes a good detector. Place one next to the amp tuned to the low end of the dial. Any difference when the amp is on or off?

I'm going to try this next thanks. Yes i have a screen resistor of 470 ohms soldered directly to socket

[Rattler66]

ithe info i currently gave was incorrect the correction is in bold

well I put in a temporary rc filter between the rect and the first node and it did nothing. it was a 60u cap going into 2 100 ohm/10 watt resistors IN PARALLEL SO MY TOTAL RESISTANCE IS 50 OHMS to my existing plate/screen node. so now I'm doubting filtering. I'm getting 3.8vac at the first node then 2.7vac at the second (this is on the cap can) .

again if I ac ground the screens/plate (both coming off the ot primary) I can get the plate/screen voltage to stabilize. if I remove the cap the meter just beeps and jumps around trying to read the voltage. so without that cap there I cannot even determine the proper bias because I can read the current (1ohm resistor at cathode) but not the plate voltage.

I'm starting to wonder if im getting any kind of rf oscillation at the screen and should I up the screen resistor to a 1k?

ALSO MY METER IS A CRAFTSMAN TRUE RMS METER AND CAN READ FREQUENCY, WILL THIS HELP?

[pdf64]

...again if I ac ground the screens/plate (both coming off the ot primary) I can get the plate/screen voltage to stabilize. if I remove the cap the meter just beeps and jumps around trying to read the voltage. so without that cap there I cannot even determine the proper bias because I can read the current (1ohm resistor at cathode) but not the plate voltage...

Just use the V HT, ie at the reservoir cap / OT CT, for the plate dissipation calculation.
Any Vdc dropped across the OT primary leg is almost certainly trivial and can be ignored.
Measure the OT primary resistance, work out the Vdc dropped, and take it off the V HT if you want to be super accurate.

[Rattler66]

Just use the V HT, ie at the reservoir cap / OT CT, for the plate dissipation calculation.
Any Vdc dropped across the OT primary leg is almost certainly trivial and can be ignored.
Measure the OT primary resistance, work out the Vdc dropped, and take it off the V HT if you want to be super accurate.

Thanks that's exactly what I'm going to do. I'm convinced it isn't a filtering issue because I'm measuring only 3.8 vac at the first node and 2.7vac at the second. That isn't bad at all. At this point if it's quietish and biased correctly then who cares. If it sounds bad that's it. I'll either eliminate the ul and go straight pentode or ditch the ot and go single ended. Thanks

[Rob's Radio-Active]

Thanks guys

[Tom Phillips]

... At this point if it's quietish and biased correctly then who cares. If it sounds bad that's it. I'll either eliminate the ul and go straight pentode or ditch the ot and go single ended. Thanks

You have provided a couple bits of evidence that point to a HF oscillation problem and you have mentioned the possibility yourself. I suggest that you address that first rather than go directly to a significant design change. Did you try using an AM radio as a sniffer as per loudthud's suggestion in post #49?

[Rattler66]

You have provided a couple bits of evidence that point to a HF oscillation problem and you have mentioned the possibility yourself. I suggest that you address that first rather than go directly to a significant design change. Did you try using an AM radio as a sniffer as per loudthud's suggestion in post #49?

Thanks no i didn't try the am radio. I forgot that one. I gotta find one first. I'll report back. Thanks