UK Vintage Radio Repair and Restoration Discussion Forum

UK Vintage Radio Repair and Restoration Discussion Forum (https://www.vintage-radio.net/forum/index.php)
-   Components and Circuits (https://www.vintage-radio.net/forum/forumdisplay.php?f=40)
-   -   Transistors for Fuzz Pedals. (https://www.vintage-radio.net/forum/showthread.php?t=139761)

Craig Sawyers 11th Sep 2017 9:21 am

Transistors for Fuzz Pedals.
 
Split from this thread:-

http://www.vintage-radio.net/forum/s...d.php?t=139659

There may be some inconsistencies.

I've been down this route, building a circuit replica of the two-transistor Arbiter Fuzz Face that Hendrix used (as a gift to my 32 year old amateur guitarist son).

But because Arbiter just shoved in any old germanium transistor they got their hands on, Hendrix bought them in quantities of 50 so he could select the very few that actually worked properly, and didn't keel over when the unit was used in hot conditions (ie when leakage shifted the bias condition enough).

For the deceptively simple circuit to work properly as originally designed, the transistors need to have a beta of 70 to 100 (most that I tested have a beta of much lower than that - 40-ish), and ideally to have a leakage of less than 100uA at 20C. And the reason that germanium, like the AC128, "sounds" better is because of their truly awful frequency response - hard fuzz doesn't have a hard biting edge.

I used a Peak DCA55 to select devices, because I'm always looking for an excuse to buy another interesting bit of test gear.

There are silicon versions out there with modified component values to try and tame the much higher beta and a peppering of capacitors to slow down the edges with hard fuzz. They sound pretty good, but very much not the same as the germanium versions. And like everything to do with musicians there are strong opinions out there on distortion pedals.

Craig

paulsherwin 11th Sep 2017 2:08 pm

Re: Transistors wanted
 
The pedal makers of the late 60s / early 70s used the cheapest functioning transistors they could find, which at that time meant Ge types as they were obsolescent and being sold off relatively cheaply. That's why you find them built with OC45s (obsolete RF) and similar oddball types.

ms660 11th Sep 2017 2:36 pm

Re: Transistors wanted
 
I made a few treble boosters for guitarists in 1965, OC44 was the specified choice on the original schematic which the first guitarist supplied scribbled on a fag packet, never forget it, was working for Fred Dawes service dept in a basement in Hale, the guitarist, who I knew, turned up at lunch time with his guitar and a small practice amp to give it a go, all was plugged in and the wick turned up, intro to Baby Please Don't Go (Them style) amazing...

Lawrence.

thomasolsen84 11th Sep 2017 3:05 pm

Re: Transistors wanted
 
I have tried a couple of silicon transistors in other circuits. Sound great. But germanium has a distict tone to it. Kind of like tube vs transistor amps. Germanium adds a warmer tobe and you can roll off the volume without it sounding harsh. A great warm tone. I measure the transistors using a transistor tester.

What i like to see is a hfe of about 120 and leakage under 300 ua. Never had any luck with npn transistors. They're usually way too much leakage to them. Making impossible to bias correctly.

Pnp transistors is way better in leakage.
Cool that you built that treble booster for him leading into thst song. I have to check it out. But I bet it sounds cool if it's like them. The best Van Morrison has done ever in my opinion.

MrBungle 11th Sep 2017 3:25 pm

Re: Transistors wanted
 
I think the bias problems are mainly because they are current biased which is more than a little unstable.

London John 11th Sep 2017 3:50 pm

Re: Transistors wanted
 
Quote:

Originally Posted by ms660 (Post 974903)
I made a few treble boosters for guitarists in 1965, OC44 was the specified choice on the original schematic.....

The first Rangemaster "treble boosters" used both OC44 and OC71.

Others, like the Hornby Skewes Treble Booster, used a 2N4061.

Both were produced in 1965.

Oldcodger 11th Sep 2017 4:45 pm

Re: Transistors wanted
 
I've got a few old books from the 60's/70's which feature FUZZ boxes and other effects. From what I read it's only a matter of driving the circuit into overload ,and generating square waves. Perhaps, OP might look at doing this with IC's , looking back at how the likes of GEC and the olde telecomms makers produced ye olde carrier generation frequencies to modulate audio into hypergroup ranges. Take one frequency ( as in take two hares) , amplify it and pass it across a coil. When the coil saturates- it generates harmonics. The harmonics were simply rectified ( to remove odd harmonics) and filtered ( to get the desired frequency) .
When the wave is changed from sine to square, there's all the harmonics wanted.

Station X 12th Sep 2017 8:22 am

Re: Transistors wanted
 
I tested a couple of OC81D's and they gave greatly different results.

Example 1. hfe 258. Leakage 0.38mA.
Example 2. hfe 61. Leakage 0.04mA.

Ed_Dinning 12th Sep 2017 8:45 am

Re: Transistors wanted
 
Hi Graham, it will depend on the tester you use, and if it measures small signal or DC gain; but the high gain on the leaky transistor could be a false reading.

Ed

Station X 12th Sep 2017 8:53 am

Re: Transistors wanted
 
Peak Atlas DCA55.

http://www.peakelec.co.uk/acatalog/jz_dca55.html

paulsherwin 12th Sep 2017 10:33 am

Re: Transistors wanted
 
The high reading is very suspicious for an OC81D. The other one is as you would expect.

Craig Sawyers 12th Sep 2017 11:46 am

Re: Transistors wanted
 
The one I built was a clone of the Arbiter http://fuzzcentral.ssguitar.com/fuzzface.php . Went the whole hog and bought replica single sided boards, carbon composition resistors etc.

The fuzz characteristic is that as the control is advanced, the signal soft clips asymmetricaly - kind of a half-wave rectified appearance. As the control is advanced it progressively clips on the positive and negative transitions. But it also depends on the drive level from the guitar. For a circuit with two transistors, four resistors, three caps and two pots you can get a whole range of effects.

Of course people analyse the thing, including spice modelling https://www.electrosmash.com/fuzz-face . Although the AC128 spice models are not correct, the waveforms are pretty close to what I measured using a signal generator for the input.

The key thing with this little circuit is that the collector of the second transistor should sit at about 50% of the rail (9V battery so about 4.5V), but anything between 30% and 70% of rail should work OK. Tweaking collector resistors to achieve that might be necessary for particular values of beta that you have. In fact some clones have a pot in the collector of the second transistor to set the bias conditions http://fuzzcentral.ssguitar.com/fuzz...facelayout.png


Craig

Argus25 12th Sep 2017 1:10 pm

Re: Transistors wanted
 
Hello Thomas,

I can let you in on a secret about germanium transistors and their applications in musical instruments such as guitars and FX pedals etc.

Most American instrument makers of guitars and pedals, if they were not scrimping on the dollar, selected the one and only one best suited transistor at the time, in the early to mid 1960's, made by RCA, which was a low noise germanium type, this was the 2N2613. That is if they didn't subs in a poorer/cheaper older part like a 2N408 or combine it with a 2N2613 as the second stage where the noise wasn't as critical.

Oddly in the USA there were few other types that were specifically low noise. In the UK there was the approximate equivalent of the 2N2613, the AC107. These transistors were initially designed and conceptualized for the input stage of Tape deck amplifiers.

Gretsch in the USA started to put them (2N2613's) in as treble boost amps in guitars like the Rally and the Jetfirebird model in the mid '60's. Gretsch lost the original documentation on these so I reverse engineered them, the analysis and history is here:

http://worldphaco.com/uploads/Gretschdoc.pdf

In the UK slightly less satisfactory transistors were selected for the treble boost application & fuzz, these were transistor radio transistors, like the OC44, OC45, OC71, OC72 etc, because these were less expensive than the AC107, at the time.

What has been neglected though is this:

The Japanese, in the early '60's era, were not confined to one low noise germanium transistor type. Because they basically pioneered battery operated portable tape recorders they made a number of types, there are at least 5. Their classics, equally as good (if not better) than the UK's AC107, or the USA's 2N2613 for this music application, these are the remarkable 2SB439 and 2SB440, both were made by Toshiba. I will let you discover the others, it makes for an interesting journey.

thomasolsen84 12th Sep 2017 2:52 pm

Re: Transistors wanted
 
Thanks a lot for great input. I really appreciate it. I have been trying to look outside the most common types of transistors, mostly because of the price and avalibility and alternative transistor suggestions are really appreciated.

Some of the russians i've tried has been looking great on the datasheet, but has too low hfe because the input current has been too low, so the transistor is not letting out enough gain because it's about 0.1 Milliamps of current on the input stage.

Thanks a lot for the japanese suggestions. I've tried a couple of different japanese 60's transistors with great results. Really good hfe range and low leakage. So I'll definitely will be on the lookout for these you mentioned.

If someone has suggestions for a low leakage NPN transistor that would be much appreciated.

Argus25 12th Sep 2017 3:14 pm

Re: Transistors wanted
 
Quote:

Originally Posted by thomasolsen84 (Post 975172)
Thanks a lot for the japanese suggestions. I've tried a couple of different japanese 60's transistors with great results. Really good hfe range and low leakage. So I'll definitely will be on the lookout for these you mentioned.

If someone has suggestions for a low leakage NPN transistor that would be much appreciated.

I have seen a lot of folks try Japanese 2SA and 2SB series transistors in fuzz & other music/pedal applications, but for some reason they have used common radio types and missed or overlooked the low noise types.

I assume you mean a Germanium NPN ? an OC139 makes a reasonable audio transistor and in the Japanese camp off hand a 2SD11 which was used in the sync separator of early transistor TV's but it would be fine for audio. You could also try an AC127.

Craig Sawyers 12th Sep 2017 4:23 pm

Re: Transistors wanted
 
Quote:

Originally Posted by thomasolsen84 (Post 975172)
Some of the russians i've tried has been looking great on the datasheet, but has too low hfe because the input current has been too low, so the transistor is not letting out enough gain because it's about 0.1 Milliamps of current on the input stage.

The original fuzz face operates with a first transistor collector current of about 0.25mA, and the second transistor with about 0.38mA for a beta of 90.

Don't know about the other fuzz circuits you are looking at

Craig

Techman 14th Sep 2017 12:56 am

Re: Transistors for Fuzz Pedals.
 
3 Attachment(s)
I bought this fuzz pedal back in the early 70s when I was attempting to be a rock guitarist in a local band. I remember replicating the circuit with the intention of building copies. I think i got to the stage of the etch resist on a trial board, but that's as far as it got. I suppose these things are becoming collectors items now.

mictester 15th Sep 2017 5:57 pm

Re: Transistors wanted
 
Quote:

Originally Posted by Craig Sawyers (Post 975192)
The original fuzz face operates with a first transistor collector current of about 0.25mA, and the second transistor with about 0.38mA for a beta of 90.

Some years ago, I got fed up with my old germanium Fuzz Face making strangled fart noises when it got warm. I came up with a fairly simple automatic bias adjustment circuit that compensated very effectively for the change in the transistor characteristics, and kept the collector voltage of the second transistor nailed to 4.6V over a very wide range of temperatures - I'll post it if anyone's interested.

The other approach I tried (successfully) was building the standard Fuzz Face circuit with silicon low gain transistors (I used 2N2369A RF devices or two of the transistors in a CA3049 transistor array IC). The trick is to bypass the base to the collector of each transistor with a small value capacitor. This gets rid of the high frequency nastiness! It's worth experimenting with the values until you get something that sounds good. 220p or 330p are good starting points.

The acid test was to take a germanium Fuzz Face and the modified silicon one to our local rehearsal studio. None of the musicians who tried them could tell the difference! :thumbsup:

Lucien Nunes 15th Sep 2017 7:01 pm

Re: Transistors for Fuzz Pedals.
 
I think the fuzz pedal is one of those peculiar situations where the reason a replica-builder chooses a component is different to the reason the original manufacturer chose it. In the days of the originals they would use any parts that could be got conveniently and cheaply, because the aim was to produce a particular effect and they knew that within reason, any one of a number of alternatives would give that effect. Now, when building a replica, replicating the original hardware is as important to people as replicating the effect, so even if the original choice was a poor one made on cost grounds, they want to make the same poor choice again.

In some cases this is compounded by a lack of knowledge of the effects of component choice on circuit performance, that leads to the well known tendency to ascribe particular characteristics to parts that they don't actually have - e.g. a 'mellow' capacitor or a 'bright' resistor. Noise is a significant nuisance in music amplification and if people would select components more on their contribution of noise, they would probably get a better subjective result as it is rare that one actually wants the noise as part of the recreated sound. Non-linearities obviously yes, because you want to re-create the original distortion spectrum, but hum and noise are invariably bad outside of musique concrète and there's usually too much noise.

AC/HL 15th Sep 2017 7:39 pm

Re: Transistors for Fuzz Pedals.
 
Three posts split to a new thread: https://www.vintage-radio.net/forum/...d.php?t=139837


All times are GMT +1. The time now is 7:39 pm.

Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.