I was wondering how I can sort out the right JFETs for this pedal. I want to use the 5.6k drain resistors like the original. Do I just pop them in and then measure? And if so, do the voltages of the JFETs interact with one-another?
Or must I build a vero testing-board and test them there?
Quote from: fixxe on November 14, 2014, 07:33:08 PM
I was wondering how I can sort out the right JFETs for this pedal. I want to use the 5.6k drain resistors like the original. Do I just pop them in and then measure? And if so, do the voltages of the JFETs interact with one-another?
Or must I build a vero testing-board and test them there?
the best thing its build the board, then socket the transistors, and then, try fets until you get 4.5V aprox on each try first on 1st stage when you get the 1st good, leave connected the transistor and move to the next and do the same for all
Chi_boy's build document has the testing procedure. You need to match the JFETs on their own first, and then if you really MUST use exactly the same resistors as the original, you need to find matched pairs that bias correctly at 5.6K. However, I can tell you it's not going to matter too much and you're going to need a giant pile of J201s to get enough pairs.
You can rough sort the FETs into pairs if the board is socketted, but the bias point of a FET on its own will not be the same as a pair, so you have to do them one at a time before trying each pair.
Yes! That was the build document I was looking for, thanks!
Hey there, I did the black 65 without matching the fets and it worked great for me. Anyway, by the timeI made that pedal I wasn't sure on how to test the jfets, but now I've found out that someone has a ready to etch board and fully documented pdf to make a jfet matcher pcb. If you want PM me and I'll send you the project. I'll be building one for myself. It's very simple, just 3 10k resistors, 1 10uf cap and a TL071.
I use the Jfet matcher that came with the Nomnom. Methinks Madbean should offer it as a separate build just for cases like this. It's easy to measure J201's with that. Giant piles of J201's for testing have become expensive though, and reputable sellers that sell them in such large numbers are both rare and a dying breed. Unless you want to use Chromosphere SMD conversion boards.
Whoever designs these boards that require biased matching JFETS should really incorporate trimpots into them. Using resistors really makes it a pain in the ass. I can understand that pro builders like Wampler don't use them as they got probably shitloads of J201's in stock and since trimpots cost more then resistors try to keep costs down, but for a cloned pedal I'd consider the extra cost for trimpots money well spent.
I bought 30 J201's, discovered several matching groups and found it far, far easier to socket the draining resistors to correctly bias them then to hope that I would a matching group of J201's that would bias correctly to the resistors. Resistors are after all dirt cheap, J201's are not.
Quote from: Muadzin on November 19, 2014, 09:34:21 AM
Whoever designs these boards that require biased matching JFETS should really incorporate trimpots into them.
+100
Quote from: Muadzin on November 19, 2014, 09:34:21 AMWhoever designs these boards that require biased matching JFETS should really incorporate trimpots into them. Using resistors really makes it a pain in the ass. I can understand that pro builders like Wampler don't use them as they got probably shitloads of J201's in stock and since trimpots cost more then resistors try to keep costs down, but for a cloned pedal I'd consider the extra cost for trimpots money well spent.
It's not just cost. For one thing, trimpots can be noisy. For another, they take up space on the PCB and sometimes the point of a project is to do it in a certain size enclosure that might be impossible with the trimpots there.
Also, FETs get matched in these pedals for reasons that go beyond "will this bias at the same voltage?" It needs to bias at the same voltage
with the same drain/source resistance while producing the same amount of gain the pedal is designed around. You could have a trimpot on the whole drain assembly, but you're still stuck matching the two FETs. (If you pair them up and one draws more current than the other, it won't function the same way.) If you just match two FETs but they bias +/- a quarter volt or so off from the fixed resistor, then you're not going to produce the amount of gain needed to make the pedal sound right. It's the same reason you use J201s instead of 2N5457s or MPF102s, because the J201 is the FET called for in
this design.
Did you miss where the OP said he wanted to use the same drain resistor as the original? Maybe that indicates that he's trying to build an actual clone.
If you don't care about doing the circuit right, then you could just throw any old thing in there, use a single J201 instead of matched pairs (you'll get about 66% the amount of gain, is that close enough?) and put the FETs on a breadboard first to figure out the closest fixed drain resistance.
Besides, you can solder two legs of a trimpot into a PCB for fixed resistors. The reverse is also true.
Quote from: midwayfair on November 19, 2014, 01:53:21 PMIt's not just cost. For one thing, trimpots can be noisy. For another, they take up space on the PCB and sometimes the point of a project is to do it in a certain size enclosure that might be impossible with the trimpots there.
Also, FETs get matched in these pedals for reasons that go beyond "will this bias at the same voltage?" It needs to bias at the same voltage with the same drain/source resistance while producing the same amount of gain the pedal is designed around. You could have a trimpot on the whole drain assembly, but you're still stuck matching the two FETs. (If you pair them up and one draws more current than the other, it won't function the same way.) If you just match two FETs but they bias +/- a quarter volt or so off from the fixed resistor, then you're not going to produce the amount of gain needed to make the pedal sound right. It's the same reason you use J201s instead of 2N5457s or MPF102s, because the J201 is the FET called for in
this design.
Did you miss where the OP said he wanted to use the same drain resistor as the original? Maybe that indicates that he's trying to build an actual clone.
If you don't care about doing the circuit right, then you could just throw any old thing in there, use a single J201 instead of matched pairs (you'll get about 66% the amount of gain, is that close enough?) and put the FETs on a breadboard first to figure out the closest fixed drain resistance.
Besides, you can solder two legs of a trimpot into a PCB for fixed resistors. The reverse is also true.
I got the part of matching FET's, in fact didn't I say
boards that require biased matching JFETS? I've build a Pinnacle and Plexi Drive with matched FET's. The FET matcher that comes with the Nonnom board was a god sent in that regard. For the Pinnacle I had to socket resistors until I found the right ones. The Plexy Drive was a Rullywow board and it was designed with trimpots. Man, that made the biasing so much more easy.
I get the point about trimpots maybe being noisy, and that may be a valid point if you're Brian Wampler, but me being me and dirtpedals being noisy anyway that's a tradeoff I'm willing to make. As for enclosure size, a cermet trimpot is not that much bigger then a single resistor in length and 3 of them next to each other . And unless you're going for a 1590A box I doubt having a couple of trimpots will take up that much more space that it won't fit into a 1590B. The Pinnacle PCB required only two drain resistors that had to be biased correctly and the Plexi Drive PCB was even smaller and still fitted 3 trimpots.
And no, I did not miss the part where wanted to use the exact same drain resistor as the original. The point I was trying to make was that its easier to change resistor values to suit a group of matched JFETS then to find a group of JFETS that will suit the resistors. Because resistors cost nothing whereas a giant pile of JFETS will cost a lot. And as long as the JFETS are matched and then correctly biased it should not change the sound one damn bit. Or am I wrong in that regard?
And yes, you can always solder two legs of a trimpot into a PCB. It will be harder though and require some creativity to get them in place if the resistors you are trying to replace are in between rows of tightly packed resistors. The reverse however is always easy. So why skimp on that?
I've started avoiding trim pots for basing jfets because the noise they add is not the kind of noise I want.
If you find them so appealing, I can appreciate that, but it sounds like a great reason to start working on your own layouts.
Quote from: Muadzin on November 20, 2014, 11:19:21 AMAnd as long as the JFETS are matched and then correctly biased it should not change the sound one damn bit. Or am I wrong in that regard?
Yes, you ARE wrong in that regard. A JFET that biases to 4.5V with only 3.9K resistance will have less gain than one that biases at 5.6K. One that biases at 10K will have significantly more. The Black 65 was designed to use paired J201s with identical gain levels at a particular bias point. The bias point matters because it defines how the wave is clipped. The gain level matters because it defines how much it will clip the next stage.
If it doesn't matter to you to build it exactly like it was intended, then use your breadboard to find pairs and their drain resistors.
I do agree that it's cost prohibitive for many people to buy enough JFETs to find suitable matched pairs for this effect, but the whole point of the PCB was to be able to build a clone of the pedal; it's not going to be designed to take shortcuts or skimp on the details.
This maybe a stupid question but I always wondered. Do large companies get parts made in the spec they need from companies like Fairchild? Can they order 10,000 units of the same spec, with in reason? or.. are large companies constantly having to re-bias circuits before shipping?
Cody
+1 with what Jon's been saying.
I've build more than a few of these types of drives and I'm at the point where, to me, getting it in the ball park is good enough and an exact clone doesn't really concern me. Actually, my favorite Black 65 I did was VERY low gain compared to stock. Sounded awesome to me.
At this point, the best route is to buy a batch of 100 mmbf201 (SMD) at the $0.14 price break and a pile of adapter boards. Being from the same batch, though, they might even be too close in tolerance to one another to get the exact ones you need.
If you absolutely HAVE TO HAVE a perfect clone, buy a genuine unit second hand.
If you're building it for someone else, tell them you modified the design a bit for a 'more complex gain structure to emphasize certain harmonics' or something like that.
Quote from: selfdestroyer on November 20, 2014, 05:45:47 PM
This maybe a stupid question but I always wondered. Do large companies get parts made in the spec they need from companies like Fairchild? Can they order 10,000 units of the same spec, with in reason? or.. are large companies constantly having to re-bias circuits before shipping?
Cody
It depends. The FET manufacturing process has pretty wide limits. I've never dealt with FET manufactures, but I have dealt with other component manufacturers that will bin them for you for a nominal fee.
Binning is a time honored practice done by either the component manufacturer or the end equipment manufacturer. My guess is that the pedal companies are doing the binning, but I don't know that for certain.
Quote from: pickdropper on November 20, 2014, 08:38:40 PM
Quote from: selfdestroyer on November 20, 2014, 05:45:47 PM
This maybe a stupid question but I always wondered. Do large companies get parts made in the spec they need from companies like Fairchild? Can they order 10,000 units of the same spec, with in reason? or.. are large companies constantly having to re-bias circuits before shipping?
Cody
It depends. The FET manufacturing process has pretty wide limits. I've never dealt with FET manufactures, but I have dealt with other component manufacturers that will bin them for you for a nominal fee.
Binning is a time honored practice done by either the component manufacturer or the end equipment manufacturer. My guess is that the pedal companies are doing the binning, but I don't know that for certain.
That makes sense. Thanks for the info.
Cody
Quote from: jubal81 on November 20, 2014, 06:23:42 PMActually, my favorite Black 65 I did was VERY low gain compared to stock. Sounded awesome to me.
When I tried the Tweed 57 at an amp show, I put the gain at 11:00 and played for a few minutes. Brian Wampler, being the nice guy he is, complimented my playing ... and then said, "I'm glad to see someone finally try it with the gain knob anywhere by max." Heh.
The gain increase with using pairs is 1.4x. You could use individual FETs for at least a couple stages and avoid the whole issue, if you don't need the whole distortion range.
Unlike the Tweed, the Black doesn't use the pairs - just singles - which makes it more suited to what I want than the '57.
In fact, I just built up another Blackbird today - rebiased for 15Volts with a Road Rage and no ceramic caps (my new thing) - and it's solidified its place as a top-3 all-time gain pedal in my list. With the added boost stage, it does it all from clean, fat boost to full-on fuzz. Recommend it to anyone and you really don't need to make an exact clone and go through a hundred JFETs to get an amazing pedal.
I love the way my sounds. I built it at first to use it as a Fender amp emulator and now I look at it as a great overdrive, one of the bests, even if it's not behaving like the original gain wise.
So let me get this straight, first priority is to get the jfets/drain resistors to match, then second to get them close to 5.6k?
Though getting close to the 5.6k is the harder part since my pile of jfets is not that big.
Before I Open a new thread. Has anyone here built the b65 and matched the fets using a jfet matcher for biasing around 4 volts with the 5.6k resistors?
If so, has anyone got the voltage readings? I have built a jfet matcher and it would be great If anyone had the readings already.
Quote from: Muadzin on November 20, 2014, 11:19:21 AM
Quote from: midwayfair on November 19, 2014, 01:53:21 PMIt's not just cost. For one thing, trimpots can be noisy. For another, they take up space on the PCB and sometimes the point of a project is to do it in a certain size enclosure that might be impossible with the trimpots there.
Also, FETs get matched in these pedals for reasons that go beyond "will this bias at the same voltage?" It needs to bias at the same voltage with the same drain/source resistance while producing the same amount of gain the pedal is designed around. You could have a trimpot on the whole drain assembly, but you're still stuck matching the two FETs. (If you pair them up and one draws more current than the other, it won't function the same way.) If you just match two FETs but they bias +/- a quarter volt or so off from the fixed resistor, then you're not going to produce the amount of gain needed to make the pedal sound right. It's the same reason you use J201s instead of 2N5457s or MPF102s, because the J201 is the FET called for in
this design.
Did you miss where the OP said he wanted to use the same drain resistor as the original? Maybe that indicates that he's trying to build an actual clone.
If you don't care about doing the circuit right, then you could just throw any old thing in there, use a single J201 instead of matched pairs (you'll get about 66% the amount of gain, is that close enough?) and put the FETs on a breadboard first to figure out the closest fixed drain resistance.
Besides, you can solder two legs of a trimpot into a PCB for fixed resistors. The reverse is also true.
I got the part of matching FET's, in fact didn't I say boards that require biased matching JFETS? I've build a Pinnacle and Plexi Drive with matched FET's. The FET matcher that comes with the Nonnom board was a god sent in that regard. For the Pinnacle I had to socket resistors until I found the right ones. The Plexy Drive was a Rullywow board and it was designed with trimpots. Man, that made the biasing so much more easy.
I get the point about trimpots maybe being noisy, and that may be a valid point if you're Brian Wampler, but me being me and dirtpedals being noisy anyway that's a tradeoff I'm willing to make. As for enclosure size, a cermet trimpot is not that much bigger then a single resistor in length and 3 of them next to each other . And unless you're going for a 1590A box I doubt having a couple of trimpots will take up that much more space that it won't fit into a 1590B. The Pinnacle PCB required only two drain resistors that had to be biased correctly and the Plexi Drive PCB was even smaller and still fitted 3 trimpots.
And no, I did not miss the part where wanted to use the exact same drain resistor as the original. The point I was trying to make was that its easier to change resistor values to suit a group of matched JFETS then to find a group of JFETS that will suit the resistors. Because resistors cost nothing whereas a giant pile of JFETS will cost a lot. And as long as the JFETS are matched and then correctly biased it should not change the sound one damn bit. Or am I wrong in that regard?
And yes, you can always solder two legs of a trimpot into a PCB. It will be harder though and require some creativity to get them in place if the resistors you are trying to replace are in between rows of tightly packed resistors. The reverse however is always easy. So why skimp on that?
Is the NonNom Jfet matcher better than Chi_boy's, and where can I find it?
Quote from: fixxe on November 30, 2014, 05:04:31 PM
Is the NonNom Jfet matcher better than Chi_boy's, and where can I find it?
Wrong kind of matching. Use Chi_Boy's or just use your breadboard or sockets for the transistors. Like Jason pointed out, the Black 65 doesn't require matched pairs, just that every FET get close to the right drain voltage with the same resistor.
Quote from: midwayfair on November 30, 2014, 09:19:20 PM
Quote from: fixxe on November 30, 2014, 05:04:31 PM
Is the NonNom Jfet matcher better than Chi_boy's, and where can I find it?
Wrong kind of matching. Use Chi_Boy's or just use your breadboard or sockets for the transistors. Like Jason pointed out, the Black 65 doesn't require matched pairs, just that every FET get close to the right drain voltage with the same resistor.
i just built the black 65 and its sounding pretty good. all my fets came in at like 7.5v so i bumped up the resistors to 15k and 16k on one to get em very close to 4.1v
so my question is why do all the resistors need to be the same value? i used 2x 15ks and 2x 16ks. should i go back and find some 201s that would work with all 15ks or all 16ks? and why is that if you dont mind explaining? (i know uou have some great expertise to share)
also i built one back in the day using the vero layout from tagboard effects and all my j201s were extremely close just off to something like 7v on the drains. i believe i just changed the 10ohm resistor in parallel with the filtering cap to bring all 4 of them down to 4v area at the same time. would there be a problem with that? tone wise? i seem to remember loving it but an improvement is always an improvement!
so sorry i missed your post saying that having the fets matched and biasing them with say all 10k resistors will give you a different amount of gain than fets that will bias with the 5.6k resistors. i cant imagine how many 201s woulf be required to mass produce a pedal with that theory. i mean all my 201s were pretty much right around 7.5v with the 5.6k resistors. finding matching 201s that bias at 5.6k?!? i dont think that would even be possible to run a production on. right? im not sure im understanding what midwayfair is saying completelym
but ive also read a report from someone that opened up there black 65 and found the bias resistors at a muvh larger value than 5.6k. i believe he said 15k. which is the value resistor that pretty much all the fairchild 201s i had needed to bias at 4v.
so theres a report of the original not using 5.6k to bias. if im even understanding correctly.
so if you got all of your 201s to bias at exactly 4v using 15k resistors then the pedal circuit would sound much different than if they all biased at 5.6k?
Ring ring... anyone? It would be much appreciated. As I have stuff soldered in now- sockets got removed when I thought I had it right before I read these posts?
Midwayfair I know youre a super guru, but what say you about brians pedals being measured with 12k bias resistors to get the 4v? I believe they were all 12k- and that brings another question is having all the bias resistors be the same value more/less important as having the voltage measured be close to the same on each drain?
if you have to get the bias resistors as close as possible to the stock valuse on all these jfets based amp in box dirt pedals wouldnt that really mean they cant even really be biased. You pretty much just have to the perfect jfets? Because if they all have to be the same value- and they should all be extremely close to the stock bias resistor value- that pretty much says you just gotta have the perfect jfets!
I know this has been covered before, but I feel like the measured 12k's in wamplers own build might be of at least some significance
Quote from: claytushaywood on December 04, 2016, 10:37:31 AM
Ring ring... anyone? It would be much appreciated. As I have stuff soldered in now- sockets got removed when I thought I had it right before I read these posts?
Midwayfair I know youre a super guru, but what say you about brians pedals being measured with 12k bias resistors to get the 4v? I believe they were all 12k- and that brings another question is having all the bias resistors be the same value more/less important as having the voltage measured be close to the same on each drain?
if you have to get the bias resistors as close as possible to the stock valuse on all these jfets based amp in box dirt pedals wouldnt that really mean they cant even really be biased. You pretty much just have to the perfect jfets? Because if they all have to be the same value- and they should all be extremely close to the stock bias resistor value- that pretty much says you just gotta have the perfect jfets!
I know this has been covered before, but I feel like the measured 12k's in wamplers own build might be of at least some significance
I think something to consider is if you're running a pedal company and you bought a bajillion J201s when they went obsolete, with the per-unit cost probably around 4 cents, it's a lot more efficient to build all the PCBs the same and have someone cherry pick FETs.
To my ears, as long as the JFET is biased somewhat close, the exact value of the drain resistor doesn't matter. I honestly can't hear it.
Quote from: jubal81 on December 04, 2016, 06:05:23 PM
I think something to consider is if you're running a pedal company and you bought a bajillion J201s when they went obsolete, with the per-unit cost probably around 4 cents, it's a lot more efficient to build all the PCBs the same and have someone cherry pick FETs.
This is it.
You can have one person sit there all day sorting JFETs by plugging them into a tester. They throw out everything say 10% off, and you probably get hundred of matches by sorting into finer bins, and then you resell the unused FETs as seconds or use them in some other product with a JFET buffer or something. Then you hand the "good" ones off to your production line who just solders them all in. No need to train or rely on the soldering team to bias individual transistors and you even get some of your money back at the end if you do it right.
As DIYers, we have the luxury of matching them more closely than they were likely matched in the pedal. Keep in mind when something's traced, it's a single unit. Maybe if you picked up a dozen of them, you'd find that 5k6 resulting in a bias of 7V or 6.5V or something. The bias point is less important than the gain factor in this design, and the gain factor is the drain resistance divided by the source resistance (plus whatever the transistor itself adds). If the design were going for a very specific waveform like the FETzer valve, it would have been designed differently. This sort of design is far more concerned about just getting above the .7V or whatever to clip a J201.
I am revisiting this- i looked for the chi_boy fet matching docs but cant find em.
I am still confused. Can someone spell this out for me? I need to match pairs of fets? (which ones are paired?) but the drain resistors need to be 5.6k or very close to 5.6k?
What parts can I change to get these to bias around the recommended 4-4.5v and keep the 5.6k drain resistors? Im so confused
Thanks so much for any help here!
The answers are all basically in this thread, but maybe a little out of order.
The Black 65 doesn't use matching pairs of JFETs. You don't need to use a FET matcher. That's most critical for phasers.
The gain of a JFET stage as used in the Black 65 is essentially gain = R_drain / R_source (because no current flows through the gate).
In order to get it to work (make sound without running out of headroom), you need to use a value of R_drain for each one so that the JFET biases at about 4-5 V at the drain.
If that means you need to use a different value of R_drain than the stock value, then the pedal may sound different from stock. But maybe it won't be a large difference (gain isn't the only influence on sound), or maybe you'll like it. I've only built one so I don't know how much variance there is.
If your JFETs don't let you bias it with the stock resistor values, then you can't really build it stock. You would need some different JFETs that had lower or higher Vp. But you can still make a working pedal if you choose the resistors that do let you bias it correctly. Heck, you could try biasing it "incorrectly" at 3 or 6 volts if you want. It might sound bad, or more distorted, but maybe it will sound awesome.
Agreed with Red. Though I've found that biasing so that the drain voltage is about 60% give me better results. I like 5.8V or so on 9V supply. When I look at it on the scope, that's where you get the most headroom and the clipping is softer and symmetrical (sounds less buzzy).