PUT (2N6027) from two BJTs?

[Fancy Lime]

Hi,

I ordered some 2N6027 for a bit of LFO experimentation but the order is delayed and may be some time. I am pondering trying to take a PUT using an NPN and PNP BJT. As far as I know, that is one of those circuits that work much better in theory than in practice. Has anyone tried this and wants to share some wisdom, like what transistor properties to look for, matching, etc?

Thanks,
Andy

[iainpunk]

in my experience, most transistors work, simulation software has a harder time to make it work that the real world.
the best thing is that you can make both a PUT and a Thyristor with BJT's quite easily.
the experience i have with it is when i build a Kick drum circuit that put out saw tooth kicks with infinite sustain (bass note) until you let go of the button.

cheers, Iain

[Vivek]

On another forum, I was helping some folks who built an EQD HUMMINGBIRD Clone and couldn't get it to work.

It uses PUT 2N6027 for it's oscillator

I first found some errors in the schematic

But after that was corrected, we found some PUT of a particular batch never worked while some PUT from another batch always worked.

Maybe some inconsistency / fake ?

[Fancy Lime]

in my experience, most transistors work, simulation software has a harder time to make it work that the real world.
the best thing is that you can make both a PUT and a Thyristor with BJT's quite easily.
the experience i have with it is when i build a Kick drum circuit that put out saw tooth kicks with infinite sustain (bass note) until you let go of the button.

cheers, Iain

Hi Iain, did you use just the two transistors in place of the PUT? There are some ideas floating around in other forums about using extra resistors, diodes or current mirrors to stabilize the makeshift PUT. What transistors did you use?

On another forum, I was helping some folks who built an EQD HUMMINGBIRD Clone and couldn't get it to work.

It uses PUT 2N6027 for it's oscillator

I first found some errors in the schematic

But after that was corrected, we found some PUT of a particular batch never worked while some PUT from another batch always worked.

Maybe some inconsistency / fake ?

Hi Vivek. Interesting, all schematics of the Hummingbird that I know use a UJT, not a PUT. Just like the original Vox Repeat Percussion. So I would not expect a 2N6027 to work at all without some modifications. Moosapotamus detailed the necessary changes in the documentation for the Skippy tremolo. Are these the errors you are talking about?

Cheers,
Andy

[Vivek]

Yes the original HUMMINGBIRD used UJT


The builders I was assisting built the version at https://pcbguitarmania.com/product/tremolo-device/

It uses a PUT

The build doc has the schematics. https://pcbguitarmania.com/wp-content/uploads/2019/06/Tremolo-device.pdf

There were some error in that schematic.


I did come across the Moosapotamus Skippy tremolo with the details of replacing a UJT with a PUT. http://moosapotamus.net/images/Skippy-v1-scheme.jpg

The PCB GUITAR MANIA circuit is a bit different.

But the interesting thing was that one batch of PUT never worked but all PUTS from another batch did !!!

[Fancy Lime]

Yes, I think one batch not working while others work sounds like a bad batch or fake items. They really do fake everything these days.

Do you mind sharing what the errors are in the PCBmania schematic? At this point, I have come across so many wrong schematics by this particular vendor that we might want to start a thread collecting their errors. That seems to be a source of great frustration for many beginners.

Cheers,
Andy

[iainpunk]

in my experience, most transistors work, simulation software has a harder time to make it work that the real world.
the best thing is that you can make both a PUT and a Thyristor with BJT's quite easily.
the experience i have with it is when i build a Kick drum circuit that put out saw tooth kicks with infinite sustain (bass note) until you let go of the button.

cheers, Iain

Hi Iain, did you use just the two transistors in place of the PUT? There are some ideas floating around in other forums about using extra resistors, diodes or current mirrors to stabilize the makeshift PUT. What transistors did you use?

im not 100% sure, but like 85%, i believe BC547b and BC557a, since i have those in abundance, but the kick enclosure is glued shut, and i'm not prying it open
no extra resistors, no extra capacitors or other stuff, most of those efforts only seem to make it less stable in my experience. but with different transistors, your mileage may vary. the only truly logical add on would be a pull up/down resistor for the only node in between the transistors that has no outside connection.
(pull down for PUT, pull up for Thyristor, i think 1M is a good starting point, if you need it, i wouldn't include it if it works fine without)

Yes the original HUMMINGBIRD used UJT

an UJT is just a less flexible and less stable PUT

cheers, Iain

[Vivek]

Do you mind sharing what the errors are in the PCBmania schematic?
Cheers,
Andy

The vendor tried to incorporate a rate indicator LED

But he placed it at a wrong place, which made gate voltage high

Then the Rate capacitor never reached the value needed to fire the PUT ie no oscillation

I removed the LED and everything started to work as expected

Then 3 builders verified this fix and confirmed it works

[Fancy Lime]

in my experience, most transistors work, simulation software has a harder time to make it work that the real world.
the best thing is that you can make both a PUT and a Thyristor with BJT's quite easily.
the experience i have with it is when i build a Kick drum circuit that put out saw tooth kicks with infinite sustain (bass note) until you let go of the button.

cheers, Iain

Hi Iain, did you use just the two transistors in place of the PUT? There are some ideas floating around in other forums about using extra resistors, diodes or current mirrors to stabilize the makeshift PUT. What transistors did you use?

im not 100% sure, but like 85%, i believe BC547b and BC557a, since i have those in abundance, but the kick enclosure is glued shut, and i'm not prying it open
no extra resistors, no extra capacitors or other stuff, most of those efforts only seem to make it less stable in my experience. but with different transistors, your mileage may vary. the only truly logical add on would be a pull up/down resistor for the only node in between the transistors that has no outside connection.
(pull down for PUT, pull up for Thyristor, i think 1M is a good starting point, if you need it, i wouldn't include it if it works fine without)

Yes the original HUMMINGBIRD used UJT

an UJT is just a less flexible and less stable PUT

cheers, Iain

So medium to low gain small signal. I think I'm going to try 2N3904/06 and BC227/37 and see what happens. Have plenty sitting around.

Do you mind sharing what the errors are in the PCBmania schematic?
Cheers,
Andy

The vendor tried to incorporate a rate indicator LED

But he placed it at a wrong place, which made gate voltage high

Then the Rate capacitor never reached the value needed to fire the PUT ie no oscillation

I removed the LED and everything started to work as expected

Then 3 builders verified this fix and confirmed it works

I was wondering how the Rate LED could possibly not mess with the oscillator. I've seen this as a mod on some forum, but cannot remember where. Makes me wonder if PBCmania just copies what they find floating around the interwebs without checking if it works. It should be easy enough to add a Rate LED via a buffering BJT.

Cheers and thanks,
Andy

[Fancy Lime]

So I tried breadboarding the Skippy oscillator PUT version with a 2N3904/06 pair in place of the PUT, feeding another NPN that pulls on an LED. No oscillation. I will check for errors and draw the circuit tomorrow. Assuming I find no errors, can someone point me to a correct schematic of how to wire up the transistors? It should be straight forward enough but the examples I found on the interwebs are all different. PNP transistors tend to be used backwards a lot for some unexplained reason. All just forum posts, though and most by people who sound really confused.

Any help appreciated,
Andy

[iainpunk]

this is a circuit i know works. its implemented in audio range, but the cap can be changed to work in LFO territory.


hope this helps, Iain

[duck_arse]

this is a circuit i know works. its implemented in audio range, but the cap can be changed to work in LFO territory.


hope this helps, Iain

I think your transistor string is upside down, maybe - the emitters should both be "outermost". I posted in the other thread .....
https://www.diystompboxes.com/smfforum/index.php?topic=126049.msg1204177#msg1204177

[iainpunk]

this is a circuit i know works. its implemented in audio range, but the cap can be changed to work in LFO territory.


hope this helps, Iain

I think your transistor string is upside down, maybe - the emitters should both be "outermost". I posted in the other thread .....
https://www.diystompboxes.com/smfforum/index.php?topic=126049.msg1204177#msg1204177

yes, that's true, i haven't double checked it. in this configuration, they won't conduct at all... they should be witched around.
also, in that oscillator, i used them in revers Beta mode, not that it made any difference tho...

cheers, Iain

[PRR]

..............maybe - the emitters should both be "outermost". ..........

IIRC an essential detail is that the "transistor action" (beta) be low enough so it craps-out at some non-zero current. Bleed resistors is one way; but reverse-connection (C and E interchanged) is another.

EDID-- as you say in other thread... 

[Fancy Lime]

So I finally got a working version:



Everything to the right of and including the C1-R2 junction is the oscillator, everything left of that is just the indicator to see if it works. The BC327-16 and BC337-16 were the lowest beta high current transistors I had. I also tried with 2N3904 and 2N3906 and it worked the same way. D3 acts as the upper limit for the charging cycle as well as the frequency indicator. R3 limits the discharge current through the transistors and D3. All in all it works quite nicely, as far as I can tell. There are two fundamental issues, though.

1) It takes a long time for the oscillator to start upon power up. How long depends on the setting. Seems to be about ten to twenty times the repetition time. Not a big deal but makes me wonder how stable this whole thing really is. The reason seems to be that the transistors start of "soft on" so that C1 charges much slower the very first time.

2) The frequency range is rather limited. R2+RV2 cannot be too low nor too high, else oscillation stops. What is too low or too high depends on the values of C1, R3, R4, as well as the supply voltage. So not really the most straight forward thing to design.

I'd say, if you want a sawtooth oscillator in the single digit Hz neighborhood and are set on avoiding ICs, this is a viable alternative to PUTs or UJTs. But I'll try to see if I can't come up with something simpler, more robust that can be swept over a wider range without switching caps.

Cheers,
Andy

[iainpunk]

in my experience, the UJT is even more finnicky than the transistors when it comes to the soft start stuff.
the saw this basic circuit creates has quite a curvy line to it, if you want to straighten that, you could add a simple current mirror or other constant current sauce.

cheers, Iain

[Fancy Lime]

Yes, it's more shark fin than saw tooth but I actually think that sounds quite good on tremolo. In fact, this shape is probably the only reason why I would use this over an opamp function generator, where getting the shark fin is a bit more work. For a real saw tooth, I'd probably go opamp.

Cheers,
Andy

[iainpunk]

an opamp shark tooth generator is really easy tho...

the coolest thing is that it has way more range with a single resistor, without cutting out when the resistance is to high or low.
the resistors can be changed to alter the response and amplitude

cheers, Iain