6BM8 + jfet reverb Hybrid Cascode recovery stage

[walkman]

Hi all,

It's has been some time since I worked on my the 6bm8 reverb that I posted here https://el34world.com/Forum/index.php?topic=30220.msg333099#msg333099

Recently I became motivated to see what more could be done to improve the reverb level and drawing from soulfetish's work
here https://music-electronics-forum.com/forum/amplification/guitar-amps/theory-design/23075-eric-barbour-s-6bm8-one-tube-reverb-imagined/page6
and here https://el34world.com/Forum/index.php?topic=27722.msg304939#msg304939

 I thought I would give something similar a go.

 The result is a stronger reverb that is near the volume as the dry signal, the long delay of choice of tank making the most audible.
 For now, I'm happy with the result (though I may look at feeding the reverb directly from the speaker as suggested by PRR).

[walkman]

Yesterday, I reworked the reverb driver section to pentode mode rather then triode strapped.

Getting their, still sounds like it could benefit from more drive, rather than more gain at the recovery.


Welcome to any suggestions...

P.S Most likely I will do  V2 version with an identical power transformer I have on hand, but in a larger enclosure to support and additional recovery gain stage.

[walkman]

Another update, I have clipped in additional resistance in parallel with the screen resistor

33k || 33k gives the strongest drive  and overall output so far and sounds quite good.
However Screen current exceeds, plate current!

I have tested it for 20 minutes or more with no obvious ill effects.
Though i don't know if 13 ma is much for the screen?

[Merlin]

Though i don't know if 13 ma is much for the screen?

Uh, something is not right there. The screen should not be pulling 13mA at 18V 

[walkman]

Though i don't know if 13 ma is much for the screen?

Uh, something is not right there. The screen should not be pulling 13mA at 18V 

Ahh thank you Merlin for the sanity check.

When I had it triode strapped, I had the plate load connected to screen pin 7 and wired through to anode pin 6.

What I have now got ( not what the schematic shows ) is a plate load of 16k5 with 0.1 if to ground and a screen resistance of 4k7 and driving the reverb from the screen.

I wonder how it’s even working at all and sounding good?

It is a little noisy but I noticed that before when it was still triode strapped, but think that it is coming from the recovery stage.

Ok too late tonight to do any more work on it

[walkman]

Wired up correctly now...

[Willabe]

Umm, there's no output jack.

[SEL49]

Umm, there's no output jack.

The TRS (Stereo) jack serves the input and output.

[Willabe]

And if you turn off the reverb where's the the dry signal?

[walkman]

And if you turn off the reverb where's the the dry signal?

The dry signal is split between the reverb send and dry inside the amp. See the detail in the top right hand corner of the schematic.

Send is from the sleeve, return is on the tip of a TRS. Dry signal is uninterrupted and passes through 470k||470k. There is a 470k mixing resistor at the output side of the reverb ( tip of TRS in the reverb unit)

[walkman]

Hi All,

 A further update, still tweaking though accept that in this configuration with the amp that the Reverb will not give cavernous dimensions, rather brings a much fuller sound and life than the dry sound alone.

[Merlin]

You have labelled pin-1 with 4V, but according to your 1M+220k resistors it should be closer to 30V. Labelling error?

Secret pro tip: You get better PSRR from a cascode if you don't include a decoupling cap on the grid. Throw away that 33uF, you don't need it!

[walkman]

You have labelled pin-1 with 4V, but according to your 1M+220k resistors it should be closer to 30V. Labelling error?

Secret pro tip: You get better PSRR from a cascode if you don't include a decoupling cap on the grid. Throw away that 33uF, you don't need it!

No it wasn’t a typo. Pin 1 and drain are close. I’m measuring the voltage with the tube in place though.

I removed the 33uf and revised the transistor to get the strongest output by ear. b+ are now similar to the previous schematic with pin 1 at 19.9 and the drain of the transistor at 19.85.

[Merlin]

No it wasn’t a typo. Pin 1 and drain are close. I’m measuring the voltage with the tube in place though.

I meant is the 220k a typo, because 4V implies a 23k resistor. With 220k the grid will be conducting forward current.

[walkman]

No it wasn’t a typo. Pin 1 and drain are close. I’m measuring the voltage with the tube in place though.

I meant is the 220k a typo, because 4V implies a 23k resistor. With 220k the grid will be conducting forward current.

When I measure the voltage on pin 1 with the tube not in place, the voltage divider functions as expected ( with ~200v supply 1M+330k gives 49v (actual value is 330k not 220k the schematic had not been updated correctly).

Are you saying that the voltage divider needs to be reworked for the this application? To prevent drawing grid current? If so can you explain to me further.

Or do I have a defective in the tube that’s drawing grid current?

It’s an old tube and I have 30+ more I can choose from and check the pin 1 voltage to see if it stays at the reference voltage of the divider.

What do you suggest?

[Willabe]

Or do I have a defective in the tube that’s drawing grid current?

It’s an old tube and I have 30+ more I can choose from and check the pin 1 voltage to see if it stays at the reference voltage of the divider.

What do you suggest?

Well, if you have 30 more, why even ask if you should try a different tube?

Pop a few different tubes in, measure the voltage like you did with the 1st tube.

Same, close to same, way different?

[Willabe]

You have labelled pin-1 with 4V, but according to your 1M+220k resistors it should be closer to 30V. Labelling error?

[walkman]

Or do I have a defective in the tube that’s drawing grid current?

It’s an old tube and I have 30+ more I can choose from and check the pin 1 voltage to see if it stays at the reference voltage of the divider.

What do you suggest?

Well, if you have 30 more, why even ask if you should try a different tube?

Pop a few different tubes in, measure the voltage like you did with the 1st tube.

Same, close to same, way different?

With my inexperience I don’t know if there is an issue with the design or the implementation.

The tubes I have are all used vintage tubes. I tested them out previously, though only graded them into strength of reverb recover by ear. When I have a chance later today I will measure the voltage on all pins with every 6BM8 that I have.

If they are drawing grid current due to wear, use, defect etc then they should not all draw the same grid current and the voltage may be noticeably different across tube depending on the grid current drawn. How noticeably I don’t know.

I can purchase a new tube and test but that will take a week to arrive and cost AUD $70! including postage for a brand new 6BM8.

[Merlin]

I don’t know if there is an issue with the design or the implementation.

Since your HT is fairly low, 30V at the grid leaves very little remaining across the tube, so it's failing to bias normally (also your JFET is rated for 40V, so you're trying to push it fairly hard with 30V at the grid). Personally I would change the 220k to 56k or so.

[walkman]

I don’t know if there is an issue with the design or the implementation.

Since your HT is fairly low, 30V at the grid leaves very little remaining across the tube, so it's failing to bias normally (also your JFET is rated for 40V, so you're trying to push it fairly hard with 30V at the grid). Personally I would change the 220k to 56k or so.

I will try that.

It’s going to be a hot day here 42°C so I may not have much opportunity till it’s cooler tomorrow.

[kagliostro]

Here 7°C and rain☹️

Franco

[walkman]

Here 7°C and rain☹️

Franco

38.8 here now and rising. Sweating🥵

Working on the reverb now. Taking some measurements at different bias voltages.

With a 68k I have 15.5v on pin 1 unloaded. 6BM8 in place strongest signal is around 8.6 on pin 1. Still getting only 0.1v difference on the drain of the LSK170 at 8.7v

Will start testing the other 6BM8.
 

[walkman]

Tested a few 6BM8 before  some how managed to kill the JFET when measuring the screen voltage ( may have shorted the pentodes anode and screen).

Have swapped in another LSK170  type (K170 GR from eBay) JFET which happened to gave an higher ldss of 5.1 vs 4.? Of the previous and rebiased. I will need to rejig the bias trimpot to get a better range.

So far no more than 0.6v difference  between pin 1 and pin 8 / drain of the Jfet on the few 6BM8 I have tested.
This may be a long process.

Back down to a much more pleasant but humid 21°C now.
Tomorrow may be my next chance for till the weekend to go much with this.

[walkman]

Bias trimpot replaced with 100ohm and 47||47 ohm to ground side giving a range of 23 to 123 ohms.

Working well now. Have found at least two tubes that result in a 1v or more difference between the pin 1 grid voltage and the pin 8 / JFET drain. The pin voltage at about 13.6 - 13.8v

Will keep slowly working through my stock to find the best ones. And update the schematic accordingly.

Another issue is a noticeable 500hz hum. It has been there all along and I haven’t yet looked at resolving it. I believe this may be a ground loop or other ground problem. At this stage all circuit ground is through a single bus wire with a couple of star points that go back to this ground bus that terminates near the input. Wonder if I need a separate ground for the first filter cap? Though I could put a 500pf before the out put jack to make a high pass filter with fc at 667hz though this may be too much bass cut

[walkman]

Found a tube that appears to bias correctly, with a bias of 1.65v with the ground to source at 115r.
(I may disconnect one of the parallel 47r to increase the range adjustment between 47r - 147r)

Updated schematic attached

[Merlin]

Another issue is a noticeable 500hz hum.

Do you really mean 500Hz? I'd assume you mean 50Hz but you Yanks have 60Hz so I'm confused!
Reverbs are notorious for humming because they so often use an SE driver (sensitive to PSU ripple), the tanks coils are sensitive to magnetic hum from the PT, and the recovery signals is so small that plenty of gain is needed (sensitive to grounding, heater supply, and PSU ripple), meaning there are several ways hum likes to creep in. Your box looks tiny, is the power transformer mounted right on it? I think I can see AC wiring passing very close to one of your audio jacks...

[walkman]

Another issue is a noticeable 500hz hum.

Do you really mean 500Hz? I'd assume you mean 50Hz but you Yanks have 60Hz so I'm confused!
Reverbs are notorious for humming because they so often use an SE driver (sensitive to PSU ripple), the tanks coils are sensitive to magnetic hum from the PT, and the recovery signals is so small that plenty of gain is needed (sensitive to grounding, heater supply, and PSU ripple), meaning there are several ways hum likes to creep in. Your box looks tiny, is the power transformer mounted right on it? I think I can see AC wiring passing very close to one of your audio jacks...

Using SpectrumView on an iPhone I   , captured the noise of the amp up full with no input (light blue) and superimposed that over a capture with the reverb on (dark blue)

There are distinct peaks at approx 75hz, 200hz, 475hz and less at 390hz. At 60hz AUS there is no difference in noise level with the reverb on or off.

Yes the power transformer is right on top of the box and the tank directly under the speaker. The box is small but is what I had on hand that would fit in back of the small combo with the tank.

[kagliostro]

Quick try to identify the problem

Use a pair of longer cables to connect the reverb tank and move it far from the actual location

What happens ?

BTW you can also disconnect the send cable from ground connection and see if something changes

Franco

[Merlin]

There are distinct peaks at approx 75hz, 200hz, 475hz and less at 390hz. At 60hz AUS there is no difference in noise level with the reverb on or off.

Check your horizontal scale! The highest peak is twice wall frequency, followed by its harmonics. Yep you have AC wiring hum, and possible transformer proximity hum.

[walkman]

Quick try to identify the problem

Use a pair of longer cables to connect the reverb tank and move it far from the actual location

What happens ?

BTW you can also disconnect the send cable from ground connection and see if something changes

Franco

Hum is no different if I move the tank, or disconnect ground to signal input of the reverb or to the output back to the amp.

The reverb on / off switch can act as audio input to the reverb recovery, the noise occurs when the tank is disconnected and the switch is open / not grounded.

 

There are distinct peaks at approx 75hz, 200hz, 475hz and less at 390hz. At 60hz AUS there is no difference in noise level with the reverb on or off.

Check your horizontal scale! The highest peak is twice wall frequency, followed by its harmonics. Yep you have AC wiring hum, and possible transformer proximity hum.

I didn’t notice the scale, though this was a quick attempt show the issue.

Will disconnect the reverb on off switch where it connects to the recovery in as the cable goes over and close to the AC as you mentioned in a previous post.

Would DC heaters help?

While the noise is noticeable at low input levels to the reverb, it is not that obvious when playing except as the reverb trails off.

This build was really a case of nothing ventured nothing gained

[kagliostro]

What I was looking for is to know if there was a ground loop due to the shield cable connection

there is an input and an output shieled cable and if the shield of the two are connected on different position a ground loop can occur

Moving the Tank in other positions was to clarify if the noise was captured because it was near a noise source (like the transformer)

----

Not sure I've understand correctly you

You say that the problem is when the tank is disconnected by the switch from the recovery

Isn't that then you have the input of the recovery stage not connected to ground and it keep and amplify noise from the circuit ?

Franco

[walkman]

What I was looking for is to know if there was a ground loop due to the shield cable connection

there is an input and an output shieled cable and if the shield of the two are connected on different position a ground loop can occur

Moving the Tank in other positions was to clarify if the noise was captured because it was near a noise source (like the transformer)

----

Not sure I've understand correctly you

You say that the problem is when the tank is disconnected by the switch from the recovery

Isn't that then you have the input of the recovery stage not connected to ground and it keep and amplify noise from the circuit ?

Franco

Yes, I tested a ground loop between the amp out and reverb in and also reverb out to amp in. I did this by making a stereo cable with the shield disconnected.

Yes I also tried it with the tank out side the combo the full length of the rca cable away from the reverb drive /recovery.

I also tested for noise from the recovery section alone with the tank disconnected. There is noise from the recovery stage. I mentioned the footswitch jack as it is effectively an audio in to the recovery that is ether open or shorted to ground.

I have now disconnected this reverb engage remote footswitch jack from where it was connected at the tank return / gate input. The noise level is similar. I have attached another overlay of the amp noise (light blue) with the reverb noise(dark blue). Not sure about the scale or if there is any way to change it. The capture from the iOS app is via the iPhone mic at about 30cm from the speaker.

Tomorrow i will remove the footswitch jack, and in its place put an on of switch and move the AC to this switch from the switch on the back of the pot. This will move the AC farther from the tube pins and reverb volume.

I also have a 2M pot floating around somewhere and may replace the pot 1M pot with that to get a stronger output signal.

Another thing I may be able to do is put together a shield around the transformer or between the transformer and the rca jacks for the tank input / return. Another shield also for the send and return signal to the reverb.