Transboost - anyone?

[lion]

The Transboost circuit was mentionen in another recent thread. I haven't seen it before and I wonder if anyone has looked into it - or built it?



Complete development thread here: https://www.talkbass.com/threads/an-inexpensive-audio-transformer-with-diy-projects.1113802/page-17

I understand the idea is to drive the small audio xfm into saturation - with some additional (and adjustable) softclip from opamp U2 with the NPN's. But I'm wondering about the circuit around U2. Looks much different from "the usual" opamp/diode topology - with the U2A plus U2B - and the NPN's and VR3.

Can anyone explain the concept and how it works. I'm also interested in your opinions.

Erik

[Gus]

Look for Musicman amp schematics. Look at the preamp section of some of them.

[PRR]

U2b is to drive the transformer, nothing for clipping.

Instead of diodes the clipper uses "Vbe multiplier" to scale the Vbe of a transistor. Two transistors like we use two diodes. Theory for the single Vbe multiplier is ample in power amplifier topics, because it is the common way to bias BJT power stages.

[lion]

Thanks. I'll have to read up on Vbe multiplier - also called "rubber diode" I see from a quick glance.

I might give the Transbooster a try when/if the Triad TY-250P gets available again.

Erik

[Vitrolin]

Thanks. I'll have to read up on Vbe multiplier - also called "rubber diode" I see from a quick glance.

I might give the Transbooster a try when/if the Triad TY-250P gets available again.

Erik

it is available but there is an extra fee from the US

http://dk.farnell.com/triad-magnetics/ty-250p/audio-transformer/dp/2280535#techDocsHook

it should be possible to find an alternative
http://dk.farnell.com/audio-transformers

[lion]

Hej Vitrolin

If you look closer you'll see that it's not available:
(Lagerstatus: Afventer levering. Vil være klar til levering d. 23-12-2016).

Anyhow, with the fee added by Farnell.dk for requiring it from their US stock it's probably better/cheaper to buy directly from US/Mouser.

Re alternatives
AFAICS the Transbooster circuit is designed and voiced specifically around the TY-250P specs. From the data sheet TY-141P seems the closest alternative, but still differs on several specs. Might be OK - but I'm not clever enough to work out exactly which details are important or even essential.

One from the Hammond serie might fit - but almost 4 times the price of the Triads.

Erik

[Vitrolin]

If you buy directly from US you might get a customs import fee...


Alternative
i would go with TY-145P it has 600R-600R impedance, which is closer to TY-250 (1k-1k) than the TY-141P (10k-10k)
i think its close enough, maybe R24 should be mounted

[lion]

Thanks Vitrolin

It just dawned on me that I have some Lunddahl xformers from some old/scrapped gear. I have both the LL7101 and LL5402, the later being an output xformer it seems.

I know Lundahl iron is in another league than the Triads and quite expensive, but could one of these maybe be put to (good) use in the Transbooster instead of just gathering dust in my parts drawer?
http://www.lundahl.se/wp-content/uploads/datasheets/7101.pdf
http://www.lundahl.se/wp-content/uploads/datasheets/5402.pdf

Erik

[PRR]

If you are going to use a "good" transformer, you may as well throw out T1 and U2b and wire it through.

The only point of the transformer here is to be "bad". Use a small bad transformer. I suspect the circuit values are deliberately scaled to the TY-250B.

[lion]

Thanks Paul

You are right of course, I'll just have to be patient and wait for the TY-250P.

Erik

[jonnyeye]

I've been slowly collecting transformers to try out for a similar purpose (transformer distortion) after I heard a friend's Diamond J-Drive, though later I found out that they claim the transformer they use (a Triad TY-142P, a 10k:2k model) is used to smooth the output, not the kind of distortion I expected.

Anyway, looking at the kind of lower impedance parts currently available, so far I've got the Mouser/Xicon 42TL016 and the Bourns LM-NP-1001. Both give good saturation on the low frequencies with limited signal inputs (takes well under a volt of signal to cause visible distortion at 100Hz), but the Mouser tx is a bit smoother in saturation than the Bourns, which gets quite mangled indeed. I haven't done any listening tests yet, just hooking things up with a signal generator and scope, so not sure which sounds better... I'm guessing that the difference is that Bourns part is a ferrite core, but I can't tell as it's potted and the datasheet tells no tales. The Mouser tx is a typical laminated stack core.

The biggest advantage of the Bourns is that it seems to be in stock everywhere, and it's cheap. If you're willing to do some experimenting, I'd add it to your list.

[D_Ex_Patria]

I've been slowly collecting transformers to try out for a similar purpose (transformer distortion) after I heard a friend's Diamond J-Drive, though later I found out that they claim the transformer they use (a Triad TY-142P, a 10k:2k model) is used to smooth the output, not the kind of distortion I expected.

Anyway, looking at the kind of lower impedance parts currently available, so far I've got the Mouser/Xicon 42TL016 and the Bourns LM-NP-1001. Both give good saturation on the low frequencies with limited signal inputs (takes well under a volt of signal to cause visible distortion at 100Hz), but the Mouser tx is a bit smoother in saturation than the Bourns, which gets quite mangled indeed. I haven't done any listening tests yet, just hooking things up with a signal generator and scope, so not sure which sounds better... I'm guessing that the difference is that Bourns part is a ferrite core, but I can't tell as it's potted and the datasheet tells no tales. The Mouser tx is a typical laminated stack core.

The biggest advantage of the Bourns is that it seems to be in stock everywhere, and it's cheap. If you're willing to do some experimenting, I'd add it to your list.

I'm building a similar project, and now I'm going to have to order both of those. I'm working with an obsolete modem transformer, and have no scope to test distortion with, so all I have to go on is my ear, which probably isn't reliable with the amount of dirt and filtering going on. Knowing they distort over 1V would put my mind at ease.

I was also thinking of snagging a TY-145P because the impedance and frequency range is similar, but it's a touch pricier.

[lion]

I've been slowly collecting transformers to try out for a similar purpose (transformer distortion) after I heard a friend's Diamond J-Drive, though later I found out that they claim the transformer they use (a Triad TY-142P, a 10k:2k model) is used to smooth the output, not the kind of distortion I expected.

Anyway, looking at the kind of lower impedance parts currently available, so far I've got the Mouser/Xicon 42TL016 and the Bourns LM-NP-1001. Both give good saturation on the low frequencies with limited signal inputs (takes well under a volt of signal to cause visible distortion at 100Hz), but the Mouser tx is a bit smoother in saturation than the Bourns, which gets quite mangled indeed. I haven't done any listening tests yet, just hooking things up with a signal generator and scope, so not sure which sounds better... I'm guessing that the difference is that Bourns part is a ferrite core, but I can't tell as it's potted and the datasheet tells no tales. The Mouser tx is a typical laminated stack core.

Thanks Johnathan. I'd be interested in the circuit you are testing the xformers in?

Erik

[jonnyeye]

I've been slowly collecting transformers to try out for a similar purpose (transformer distortion) after I heard a friend's Diamond J-Drive, though later I found out that they claim the transformer they use (a Triad TY-142P, a 10k:2k model) is used to smooth the output, not the kind of distortion I expected.

Anyway, looking at the kind of lower impedance parts currently available, so far I've got the Mouser/Xicon 42TL016 and the Bourns LM-NP-1001. Both give good saturation on the low frequencies with limited signal inputs (takes well under a volt of signal to cause visible distortion at 100Hz), but the Mouser tx is a bit smoother in saturation than the Bourns, which gets quite mangled indeed. I haven't done any listening tests yet, just hooking things up with a signal generator and scope, so not sure which sounds better... I'm guessing that the difference is that Bourns part is a ferrite core, but I can't tell as it's potted and the datasheet tells no tales. The Mouser tx is a typical laminated stack core.

Thanks Johnathan. I'd be interested in the circuit you are testing the xformers in?

Erik

I'm using a signal generator (Heathkit IG-18) fed directly into the transformer primary, and then looking at the output of the secondary with an oscilloscope... I haven't got as far as an actual effect just yet. One thing to note about my signal generator is that the output always has a minimum of 150 ohms series resistance (it's an old and fairly primitive piece of gear; all discrete transistors inside!) which tends to increase transformer distortion somewhat. As far as an effect goes, the transboost schematic looks pretty good as a platform to start with, at least in as much as it will let you hear primarily transformer distortion when the clipping threshold is set high.

Give me a couple of days and I'll take a thousand words worth of pictures of various signal levels and frequencies going through some of the small transformers I have.

[lion]

Give me a couple of days and I'll take a thousand words worth of pictures of various signal levels and frequencies going through some of the small transformers I have.

I'll follow your experiments and anything you can share with interest, thanks.

Erik

[D_Ex_Patria]

Give me a couple of days and I'll take a thousand words worth of pictures of various signal levels and frequencies going through some of the small transformers I have.

I'll also be watching for updates very eagerly. The information you already posted was extremely helpful. Thanks!

[Vitrolin]

i also find this interesting, but i think you should have some load on the secondary, to avoid large spikes that would not apear i a real circuit.

[jonnyeye]

OK, these last couple of days took longer than I'd hoped! Here's a first pass at some more information for you all. (Many pictures ahead!) Vitrolin, I'll get to your concern later on.

I've started with these three:

The Bourns LM-NP-1001, Xicon 42TL016 and Hammond 140Q. All three are nominally 600:600 ohm transformers. The first two are cheap and readily available; the Hammond has too much fidelity and is too expensive to use for distortion-as-effect (as we'll see!) but I had it lying around and figured it could be used as a control of sorts.

Here's a little bit of information about the process, so you can tell how much salt you should expect to ingest along with the information:
My test gear is old and in need of an overhaul, so don't worry too much if you can see vertical asymmetry (2nd harmonic) as I think this is an artifact of the 'scope, but can't be sure. There is always some resistance in series with the output of the signal generator, which increases the distortion (as noted before, this is fine in this application), so all of these transformers will be cleaner when driven by a proper low impedance source. The voltages chosen may seem random; they're the maximum output at each of the voltage selector steps on the signal generator - they are not the expected values (eg 250mV should be 100mV!), but the resistors in the divider are out of spec - did I say my test gear is in need of an overhaul? But they are indicative; between 2V and 3VRMS is about the most you can get out of a single opamp stage with a 9V supply, although the transboost uses a trick to get twice that much headroom, so can push the transformers a bit harder than shown here. The output voltage is tested stable, and the frequency should be stable as well.  Also, the timebase and vertical gain on my oscilloscope is entirely uncalibrated, so I did not touch the settings on each given frequency and signal level (except perhaps vertical position) in order to show relative insertion loss, etc. Finally, the loading in these photos is the scope's 1M input impedance only.
What to look for are trends and comparisons. My notes follow each set of photos. (Also, in a couple of places I forgot to change the tags so I edited them to be correct digitally. Just in case you notice, you'll know why)

First we'll start with an 80Hz tone. This is basically a low E on guitar, and in the middle of the range of a bass.
80 Hz, 250mVRMS in:

There's not too much to say between the Hammond and the Bourns here - very similar level and no distortion evident. The Xicon is already showing distress; a small but visible amount of saturation (see the slope of the rising side of the waveform?) and significant insertion loss (guestimate 6dB).

80 Hz, 650mVRMS in:

The Hammond hums merrily along, while the Bourns part is showing some saturation and the Xicon is a bit more bent than at 250mV, but not greatly so.

80 Hz, 2VRMS in:

Oh my, poor Bourns... gross saturation. Given how soon the Xicon started distorting, it's notable that it holds up far better here... Old Hammond is steady as a rock.

80 Hz, 3VRMS in:

The Hammond is finally starting to show some (albeit gentle) distortion. The other two continue on their trend towards ever more bentness - the Xicon slowly, the Bourns dramatically.

Here I've rescaled the vertical gain on the scope (with the same 80Hz, 3V input) to better compare the shape of the waveform of the Xicon with the other two:



OK, what if we change the frequency? 240Hz is (almost) the high E string on guitar, and just off most basses fretboards.
240 Hz, 3VRMS in:

The distortion is massively reduced in all three transformers at this frequency. It appears that there is less insertion loss here in the Xicon than at 80 Hz?

And just as a test:
1k Hz, 3VRMS in:

No visible distortion on any of the three. Also, the level of the Xicon is nearly up to that of the other two; some of the insertion loss is certainly frequency dependent.

Now back to Vitrolin's comment: There are no spikes from transformer saturation, so if the input is a smooth sine wave, we'll have no issues. But if the transformer is driven by a wave that's clipped square, there could be some ringing. The transboost loads the transformer with 1k||10k, or 909 ohms. Let's see what the ringing might be with various loads, starting with the 1M input impedance of the scope.  (I've used a 10kHz square wave at 1Vpk for this test to make it easier to see the ringing. It's still present at lower frequencies)

10kHZ, 1V Squarewave, 1M load (just the 'scope):

Looks like a doorbell factory. Let's try taming this by loading the secondary with a resistor; first try 10k?


10kHZ, 1V Squarewave, 10k load:

This is interesting - the Bourns is well damped here while the other two still ring, albeit less so. What about a heavier load, 1k (very close to the transboost's value)?

10kHZ, 1V Squarewave, 1k load:

The Hammond and Xicon look right, but the Bourns is dreadfully overdamped. I do wonder about the rating of the Bourns as a 600:600 ohm transformer; it seems fine as far as insertion loss goes but doesn't handle that kind of loading nearly as well. There will be notable treble loss with a 600 ohm loading (which may be fine, and could even be usable as an effect!), but the full frequency response is available when loading it more lightly.

Your comments are gladly welcomed. If you have more tests you'd like me to run, let me know, though my turnaround time will be slow. Hopefully this gives you a better picture of what to expect from various types of transformers in this type of circuit.

[D_Ex_Patria]

 

Thanks  jonny, that was fantastic! I'm glad to hear that they provide distortion at those RMS values; it bodes well for my project, as I'm putting them after an 386, which should be pushing out 10V p-p at the 12 Volts I'm powering it with.

I've ordered the Bourns and a Triad TY-145P. I might consider going to Mouser for the Xicon, but I'm trying to avoid paying their flat rate shipping.

[lion]

Thanks Johnatan - for taking the time to document your experiments and putting it all up here.

Interesting how the two cheap xformers perform so differently. I wonder if the difference could be read/foreseen from their specifications? Likewise, is the Hammond specs looking so much better. I'll need to study it more, but I'd guess not.

Erik