Joyo - JF-02 Ultimate Drive "2010" MISTAKE  [schematic]

All about modern commercial stompbox circuits from Electro Harmonix over MXR, Boss and Ibanez into the nineties.
User avatar
howmuch
Breadboard Brother
Information
Posts: 68
Joined: 09 Nov 2014, 18:31
Has thanked: 21 times
Been thanked: 58 times

Post by howmuch »

I haven't seen this pointed out anywhere, but there is a mistake in the "2010" Joyo Ultimate Drive JF-02.
I recently bought one of these (October 2014) and thought I'd check out the circuit.
The version I have is the newer one - PCB reads "2010.11.24" compared to "Nov.2009" on the older version.
"2010.11.24" PCB version
"2010.11.24" PCB version
The mistake they've made on this version is that the Ge diode D2 is the wrong way round. The "2009" version looks OK from the photos I've seen.
To be more precise, the PCB silkscreen for D2 is printed in the wrong orientation and D2 is fitted according to the (incorrect) silkscreen. So the problem is in the PCB layout/silkscreen, not the assembly.
Here's the original/fixed clipping
Here's the original/fixed clipping
D2 is preventing itself and the Q3 MOSFET that it's in series with from doing any signal clipping.
As a result, the positive signal is clipped by the single Q2 MOSFET body diode at about 0.8V, but the negative signal is only clipped by the opamp when it hits near-rail giving a very asymmetrical waveform.
While some people might actually prefer this sound ("more dynamic"), it is obviously not intended to be like this.
(BTW, I have to say that I've never owned the pedal that this is cloned from, so maybe Joyo has copied it correctly, and the original is the same!)

Here's what the signal looks like as the gain is turned up.
First the positive signal is clipped by the single Q2 MOSFET at about 0.8V.
First the positive signal is clipped by the single Q2 MOSFET at about 0.8V
First the positive signal is clipped by the single Q2 MOSFET at about 0.8V
Signal01.jpg (15.36 KiB) Viewed 13443 times
The negative signal is not clipped until the opamp starts clipping it.
The negative signal is not clipped...
The negative signal is not clipped...
Signal02.jpg (15.47 KiB) Viewed 13443 times
The negative signal is not clipped...
The negative signal is not clipped...
Signal03.jpg (15.05 KiB) Viewed 13443 times
...until the opamp starts clipping it
...until the opamp starts clipping it
Signal04.jpg (15.01 KiB) Viewed 13443 times
How to fix it
The PCB silkscreen shows D2's legend with the anode closest to the H/L switch and D2 is fitted like that. It needs to be turned around so that D2's polarity is opposite to the silkscreen, with the cathode (black line) closest to the switch.
The circuit will then give slightly asymmetrical positive MOSFET clipping and negative (MOSFET + Ge) clipping.
This is how the original D2 looked.
This is how the original D2 looked
This is how the original D2 looked
This is how the original D2 looked
This is how the original D2 looked
And this is how my D2 looks when fixed.
And this is how my D2 looks when fixed
And this is how my D2 looks when fixed
If you don't want to attempt removing D2 to turn it around, you could just add another diode across it in the opposite electrical direction.
And if you prefer symmetrical clipping, you could just link across D2.


NOTE: I accept that I could have a one-off and Joyo now fit D2 opposite to the silkscreen, making the circuit correct, although considering the PCB says "2010" and it's now 2014, I think it's unlikely.

NOTE: The MOSFET clipping is done by the body diodes in this pedal's circuit configuration, not the main channel, as some people seem to think.

NOTE: The "2010" version has an input buffer using a BS170 MOSFET. I've read that this extra buffer was added at some point to eliminate "popping". The on/off switching is done after this buffer which only passes signal when powered, so it is not True Bypass, and to be fair, it only says "Bypass" on the pedal, although I still see it advertised as True Bypass.
R.I.P. Burke Shelley, from one of my favourite (and much underrated) bands - Budgie
Their songs have been covered by Van Halen, Iron Maiden, Soundgarden, Metallica and more.
If you don't know how HEAVY this band was in the 70s, check Breadfan out.

User avatar
HeavyXIII
Information
Posts: 4
Joined: 19 Nov 2014, 21:45

Post by HeavyXIII »

I was interesting in doing some asymmetrical clipping mods to this earlier, are you saying it would be MORE symmetrical if I corrected it?

As an aside, would it sound more like a fuzz if I removed the Q2 mosfet and corrected D2?

User avatar
howmuch
Breadboard Brother
Information
Posts: 68
Joined: 09 Nov 2014, 18:31
Has thanked: 21 times
Been thanked: 58 times

Post by howmuch »

HeavyXIII wrote:I was interesting in doing some asymmetrical clipping mods to this earlier, are you saying it would be MORE symmetrical if I corrected it?
Yes, it would be closer to being symmetrical, but still asymmetrical.
If you want more asymmetry, you could stick with the "incorrect" circuit and power it from a higher voltage, e.g. 18V. The pedal looks to be safe to use at that voltage.
It's generally accepted though that a certain level of asymmetry can sound good, but too much is not so good. But if YOU like it, then who can argue? :)
HeavyXIII wrote:As an aside, would it sound more like a fuzz if I removed the Q2 mosfet and corrected D2?
I don't think so. Fuzz pedals are cheap enough if that's what you want.
R.I.P. Burke Shelley, from one of my favourite (and much underrated) bands - Budgie
Their songs have been covered by Van Halen, Iron Maiden, Soundgarden, Metallica and more.
If you don't know how HEAVY this band was in the 70s, check Breadfan out.

User avatar
RedHerring
Information
Posts: 1
Joined: 31 Oct 2014, 11:59

Post by RedHerring »

I just did this mod to my JF-02 that I purchased a few weeks ago, which was exactly like yours was, board dates and all.

Sounds pretty crunchy now. Usable actually, though I have better pedals for that job. Thanks for the heads up.

While I was in there I reflowed all the solder joints with good old kester. Some of the joints looked marginal.

User avatar
metakel
Information
Posts: 5
Joined: 26 Dec 2014, 11:25
Has thanked: 4 times
Been thanked: 1 time

Post by metakel »

But I have found an OCD schematics on the web:

Image

which shows the diode is in the same "incorrect" direction. What do you think?

User avatar
howmuch
Breadboard Brother
Information
Posts: 68
Joined: 09 Nov 2014, 18:31
Has thanked: 21 times
Been thanked: 58 times

Post by howmuch »

metakel wrote:But I have found an OCD schematics on the web:
which shows the diode is in the same "incorrect" direction.
In a way, yes...
In a way, no...
Look again at how the diode connects to the MOSFET.
If you want to leave the diode as it is, then you need to turn both MOSFETs around instead.
It's far easier to turn the diode around - 2 legs vs 6.
R.I.P. Burke Shelley, from one of my favourite (and much underrated) bands - Budgie
Their songs have been covered by Van Halen, Iron Maiden, Soundgarden, Metallica and more.
If you don't know how HEAVY this band was in the 70s, check Breadfan out.

User avatar
metakel
Information
Posts: 5
Joined: 26 Dec 2014, 11:25
Has thanked: 4 times
Been thanked: 1 time

Post by metakel »

In a way, yes...
In a way, no...
Look again at how the diode connects to the MOSFET.
If you want to leave the diode as it is, then you need to turn both MOSFETs around instead.
It's far easier to turn the diode around - 2 legs vs 6.
Alright, I missed the direction of the FETs.

By the way, I have modified mine using your method, thank you!!

Did you read again the waveform from your oscilloscope after modification? Does it show more symmetric form?
Also, I wonder why they connect both FETs this way (shorting the drain and gate).

User avatar
howmuch
Breadboard Brother
Information
Posts: 68
Joined: 09 Nov 2014, 18:31
Has thanked: 21 times
Been thanked: 58 times

Post by howmuch »

metakel wrote:Did you read again the waveform from your oscilloscope after modification? Does it show more symmetric form?
Yes I did, but did not take any photos.
It was exactly as expected - clipping the signal at roughly +0.8V and -1.2V.
metakel wrote:I wonder why they connect both FETs this way (shorting the drain and gate).
You also missed the fact that in the partial schematic I gave, the gates are connected to the sources.
When the clipping circuit is arranged like this, it makes no discernable difference to the sound whether the gates are tied to the sources or drains.

And IF they were connected gate to drain (which they're not)...
The MOSFET clipping would still be done by the body diodes, not the main channel. Each clipping leg will prevent the other leg's MOSFET reaching its gate threshold voltage.
Although, having said that, the single MOSFET Q2 would be drawing very small current at a Vgs of 1.2V. This would be imperceptible.
When the gates are tied to the sources (as in the Joyo), Vgs is always 0V.

BTW, I've seen a lot of schematics and layouts of the OCD with the MOSFET gates connected to the drains.
I've never seen any evidence that that is the way they're connected. The photos I've seen never quite show enough detail.
It could be a myth originating from the assumption that because MOSFETs are used then they are "soft clipping", which means gate-drain.
In this clipping arrangement, the body diodes dominate. The circuit would require another diode to be "soft clipping".
Can someone point to photos that clearly show the OCD's clipping MOSFETs and PCB traces, proving without a doubt how they are connected?
R.I.P. Burke Shelley, from one of my favourite (and much underrated) bands - Budgie
Their songs have been covered by Van Halen, Iron Maiden, Soundgarden, Metallica and more.
If you don't know how HEAVY this band was in the 70s, check Breadfan out.

User avatar
nooneknows
Resistor Ronker
Information
Posts: 338
Joined: 26 Mar 2008, 21:53
Has thanked: 58 times
Been thanked: 14 times

Post by nooneknows »

I can't find the 'thank you' button anymore but, thank you anyway

User avatar
mr_book
Information
Posts: 1
Joined: 09 Jan 2016, 00:44
Has thanked: 1 time
Been thanked: 1 time

Post by mr_book »

I only registered for this forum to thank you.

I had this pedal and a friend of mine gave me a real fulltone OCD to A-B test and I found the Joyo very harsh sounding. This is why i started looking for a modification and I found your post.

I bought my pedal a month ago (December 2015) and i have the exact same version. The diode was indeed reversed and no current was passing through.
I flipped the diode and WOW: I have to say that the sound has improved dramatically. This is how it was supposed to sound (much much better).

I A-B tested it with the OCD and you can now always find settings that they sound completely identical.

I also put a switch to bypass the diode to produce completely symmetrical clipping but i don't like it that much. its too smooth without any colouring.
In other words
-> Joyo with mistake (as it is sold since 2010 :/) -> clipping only one side (extremely asymmetrical) -> really harsh
-> Joyo corrected by flipping the diod (as it should be in the first place) -> slightly asymmetrical clipping due to the diod -> best sounding. Almost identical to OCD. WoW.
-> Joyo with diode bypassed (i put a switch to do it when i feel like) -> small change, sounds smoother with less bite, nice for chords , less character. I slightly prefer the corrected diod.

I cannot believe they are selling a buggy pedal for 5 years now :/

howmuch: thank you again! You need to post this everywhere so people know!
If you still have the oscilloscope and some time, can you produce some pictures for the corrected version ? I am curious about how the clipping looks for the corrected version.

User avatar
amonte
Information
Posts: 16
Joined: 25 Mar 2010, 20:25
Has thanked: 3 times
Been thanked: 10 times

Post by amonte »

Hey, just wanted to post a follow-up on my experience with this pedal. I have the same PCB version and bought my pedal at some point last year. When I took it apart today, the diode was going the correct way. So it seems they did (at some point) figure out the issue.

User avatar
bluegrasser12345
Information
Posts: 4
Joined: 27 Jun 2016, 20:54
Has thanked: 3 times

Post by bluegrasser12345 »

Thanks everyone for the great observations and info!

Fellow Joyo UD Fan and novice modder here trying to understand the concepts here. I have a UD bought in August 2015 that appears to have the exact same board and "incorrect" diode orientation as discussed here. After a little research, it appears that this diode only showed up in version 4 of the OCD (which this pedal attempted to clone). Folks who dislike the asymmetrical clipping it introduced have proposed removing the diode and bridging the holes, returning the pedal to symmetrical mosfet clipping used in previous and subsequent versions.

Others simply removed the diode from the OCD and left the holes open - wouldn't this essentially create the same scenario as this incorrect Joyo, with severely asymmetrical mosfet clipping on only one side of the wave?

User avatar
idy
Breadboard Brother
Information
Posts: 126
Joined: 14 Dec 2011, 04:41
Completed builds: FFs, Octavias, TBMKIIIs, Buzzarounds/alikes, Axis Fuzz', CS Overdriver, BMPs, Black '65s, MOSFETODs, WiiO, RAH, Klons, AC30, DLS, TSs and many variants, Small Clones, UniVibes, E. Mistress, Mutron Env Filter, Ross/Dynacomps,, EA Trems, TremLunes, Katana Boosters, BYOC Tri Boosters, other boosters and buffers, PT2399 delays, SKRM based delays, Belton Brick Reverbs, 3band parametric EQ,
Location: Idyllwild, CA
Has thanked: 27 times
Been thanked: 31 times

Post by idy »

If you remove the diode the single remaining mosfet clips in one direction as a silicone diode (.6v), in the other as a mosfet, more like 1.8v. A good assymetric clip. Very nice sound.

Which is probably what you get with the "fixed" version, since when you have two "clippers" in parallel the lower value is the only one that works. The original "Body diode blocked" mosfet would only have conducted in the direction of the other mosfet's body diode (.6), but at 1.8v and therefore irrelevant.

The way that has been suggested as a fix (reversing the diode) gives you diode in series with a body diode, so about 1.2v, in parallel with the other transistor's mosfet function, making it irrelevant. It gives you one diode drop in one direction and two in the other.

A lot of the original "mosfet" overdrives are like this, the guys who did it didn't think it through and could have gotten the exact same result without any mosfets. Which really do clip nicely and really do sound different from silicone diodes if used correctly.

User avatar
bluegrasser12345
Information
Posts: 4
Joined: 27 Jun 2016, 20:54
Has thanked: 3 times

Post by bluegrasser12345 »

idy wrote:If you remove the diode the single remaining mosfet clips in one direction as a silicone diode (.6v), in the other as a mosfet, more like 1.8v. A good assymetric clip. Very nice sound.

Which is probably what you get with the "fixed" version, since when you have two "clippers" in parallel the lower value is the only one that works. The original "Body diode blocked" mosfet would only have conducted in the direction of the other mosfet's body diode (.6), but at 1.8v and therefore irrelevant.

The way that has been suggested as a fix (reversing the diode) gives you diode in series with a body diode, so about 1.2v, in parallel with the other transistor's mosfet function, making it irrelevant. It gives you one diode drop in one direction and two in the other.

A lot of the original "mosfet" overdrives are like this, the guys who did it didn't think it through and could have gotten the exact same result without any mosfets. Which really do clip nicely and really do sound different from silicone diodes if used correctly.
Thanks IDY! So the only essential difference between correcting the diode orientation ("fixing" in this case) or removing it is mosfet+germanium or mosfet+silicone clipping, with removing it being a more extreme silicone dominated asymmetric clip? And leaving the GE diode in the wrong orientation and removing it would yield the same result?

User avatar
jbru8080
Information
Posts: 1
Joined: 14 Oct 2016, 14:03

Post by jbru8080 »

I registered just like someone else did to say Thank You!
I always liked my Joyo UD, but never loved it. Flipping the Diode has made enough of a difference to start enjoying this pedal again.
After coming across this discussion, I became suspicious that my Joyo Ultimate Drive fell into into this dilemma. Although I could never put my finger on it, I always knew that something was not quite right with my Joyo, after ripping mine apart, I confirmed it pretty quickly.
The pedal now "crunches" correctly, and is slightly brighter, although that could be psychosomatic. I actually soldered the diode back to the underside of the pcb, in case I didn't like it, and so I didn't have to rip my whiole pedal apart gain to re-fix it. But, luckily that was not needed. SO thank you very much to the person who discovered this... just fyi for those still debating fixing their Joyo, Do it, it absolutely makes a difference in a good way! Also if you google image the Harley Benton Ultimate Drive Schematic (same as Joyo), and than look a schematic of the Fulltone OCD v4, and the backwards diode is even shown on the Harley Benton (Joyo) Schematic of this pedal.. so that re-confirmed it for me. Thanks Again!

User avatar
N72
Information
Posts: 1
Joined: 22 Oct 2016, 10:28
Been thanked: 1 time

Post by N72 »

Thanks for the thread and the detailed instructions! I would share my experiences with the community.
There is a re-branded Joyo UD pedal called Ammoon AP-02 Ultimate Drive. It has simple black paint and inside the infamous Joyo printed circuit board with the "2010.11.24" sign. And the reversed diode.
These pedals can ordered from ebay for 25-30 USD with free shipping. With the known fix and a little (also known) modification it is now the cheapest solution to get "that" creamy sound.

All you need is:
- small signal Ge or Schottky diode
- 330k resistor

Since Ge diodes become hard to find and expensive nowdays, I tried a Schottky diode with success. Schottky diodes has lower opening voltages than traditional Si diodes. The unit I had in the shelf has a measured opening voltage 450mV - perfect. The characteristic differences from the Ge is not really relevant in this application since the MOSFET soft opening defines the overall sound, the diode just makes some asymmetry.

It is a bit large sized diode I found at home -- you can use any smaller Schottky diode.
It can be soldered to the PCB side with its mark facing to the switch.

The 330k resistor is soldered to the Gain pot, the two outer legs, also the left and middle legs are shorted with a little extra tin. This way there is no pot leg left "floating", now it works as a two-connector variable resistor. The effect is to strech the useful sonic change to the whole knob rotation, not just the lower 1/4.

I attached a photo via postimage, I hope it works.
Image
Thanks again for the tips!

User avatar
alexradium
Resistor Ronker
Information
Posts: 313
Joined: 31 Jan 2010, 13:31
Has thanked: 64 times
Been thanked: 68 times

Post by alexradium »

i got a brand new UD like this and looking to the circuit,the diode is correctly installed,mosfets have to be reversed,layout is wrong because gate and drain should be shortcircuit and source has to be alone.
I'll report back when i make the conversion

User avatar
alexradium
Resistor Ronker
Information
Posts: 313
Joined: 31 Jan 2010, 13:31
Has thanked: 64 times
Been thanked: 68 times

Post by alexradium »

i can confirm all of the above,the mosfets need to be reversed,not the diode,now the pedal is bright not muffled and fuzzy,that was due to the main opamp clipping alone.
Apart from the sloppy placing and quality of the components,for the price,you can get a very good sounding OD/distrtion that can get pretty much all the classic Marshall sounds.,

User avatar
Intripped
Cap Cooler
Information
Posts: 670
Joined: 25 Aug 2009, 01:03
Location: Italy
Has thanked: 989 times
Been thanked: 305 times

Post by Intripped »

In the OCD circuit the 2 mosfets are connected as in this pedal: S+G and D alone.
I'm 100% sure about this because I've examined an original OCD.
https://www.freestompboxes.org/viewtopic ... &start=520

User avatar
alexradium
Resistor Ronker
Information
Posts: 313
Joined: 31 Jan 2010, 13:31
Has thanked: 64 times
Been thanked: 68 times

Post by alexradium »

Intripped wrote:In the OCD circuit the 2 mosfets are connected as in this pedal: S+G and D alone.
I'm 100% sure about this because I've examined an original OCD.
https://www.freestompboxes.org/viewtopic ... &start=520
thanks for the tip,i was used to AMZ examples which showed the other way,plus i did a clone with a Madbean pcb which was also reversed,others make the pcb correctly but the schematic is opposite.
By the way,it is my understanding that this method uses the internal silicon diode between source and drain,no big deal compared to normal diodes...

Post Reply