The open source Jupiter-8-4-ALL project

[dikkietrom]

the placement is carefully replicated on the original boardsize, it is one voice. There are 3 drawing of the board in the service notes and I have a decent image of the board. The scheme and the layout had a few things missing but with everything combined it was possible to find everything. On the orignal it is 2 times the same circuit.

[ACME]

I also built myself a Module Board. This is only one Part of my JP8 replica.. All ICs in the filters (HPF /LPF) are original from Roland. The other BA662 are also clones from the openmusiclabs schematic. For this I have designed my own pcb.

Here comes my last copy.:

[StillNotWorking]

Very nice board ACME. From your experience are the any of the trimmers that rather should been multiturn to ease adjustments?

[ACME]

No, I dont think so. Before I decided to design circuit boards, I built the entire VCO on a breadorad, with multiturn trimmers. The setting of the VCO was very cumbersome and took a lot of time, because you will screw and screw and screw until due to reach the right value. The fine-tuning may be easier then, but is not necessary because it goes just as well with good quality single-turn trimmers and a calm hand. The last few cents are corrected by the autotune.

[dikkietrom]

I also built myself a Module Board. This is only one Part of my JP8 replica.. All ICs in the filters (HPF /LPF) are original from Roland. The other BA662 are also clones from the openmusiclabs schematic. For this I have designed my own pcb.

Here comes my last copy.:

wow that looks stunningly accurate, brilliant! Can I ask you some questions? I have a lot

[dikkietrom]

ACME you might have a ground loop down left. On the original board there is no connection. Working on the zones now.

[boops]

I can help with the IR3R01 pm me. Let me know where are you

I am somewhat surprised and confused about the dac's, I always thought memory was 12bit or 14bit but it is 8bit on the controller for pot data and 12/14bit on the interface for key cv. The 8 bit is a discrete ladder.

Btw it is not 32 sh but "only" 23 per dac throug 3x8-channel multiplexers, I have no clue about the s&h refresh rate, the z80 has 8mhz clock.

Anyways the voice scheme is annotated with original references now, the original scheme has some missing references but that can be recalled from the board layout. Found one more wrong junction in the kicad scheme. In hindsight it was not so bad that I had to go over it a few times cause I found mistakes that otherwise might have been unnoticed.
I am going to take a little break. Looking forward to replicate the component placement.

[dikkietrom]

That would be great! I was seriously considering buying an mks7

[ACME]

wow that looks stunningly accurate, brilliant! Can I ask you some questions? I have a lot

Feel free to ask me, I also have a lot of questions.

Markus

[dikkietrom]

wow that looks stunningly accurate, brilliant! Can I ask you some questions? I have a lot

Feel free to ask me, I also have a lot of questions.

Markus

Ok cool, can you verify my resistor analysis


There are 3 types on the board

- R25G, user groove suggested carbon 1/4 watt 2% On board pics I see red band at the location so I go with carbon 1/4 watt 2% as suggested.

- Carbon R25 , carbon 1/4 watt, gold band on pics > carbon 1/4 watt 5%

- Metal Film RB25FX, 1/4 watt, brown band > MF 1/4 watt 1%

Did you came to the same conclusion?

[dikkietrom]

It is raining juptiter replica projects, nice! Just discovered the other project here, started roughly at the same time, check out the first post dates. It is in the air.

removed the zones and went with 2mm tracks to get things done. For ground and +-15v, in the spirit of. Let the rest go all night, see what tomorrow brings.

[dikkietrom]

proto beta

[dikkietrom]

Module controller scheme is at about 80% replicated for all CV output, not much to show for now. I am getting more and more intrigued by the Jupiter 4 along the way. Since the adsr, vcf and vca will have the same core components and design it is going to be a mix of both. JP4 seems to have the same vco conceptually but less stable I guess, maybe better for the ears? Jupiter 4 is by far what comes the most close to a JP8 and vica versa. Jupiter 6 and mks80's are too far off.

Remarkable that from most synths the users always want the earlier version, also for the JP4 while the first version has the discrete ba662 OTA vs later the IR3109. Both synths rely heavily on the ba662, the accuracy of the ba662 clone will determine for a great part the resemblance.

[Pesca]

Whats your plan for the envelopes? Clone the ir3r01 or stick with the Tom Wiltshire ADSR? Just curious

[dikkietrom]

5 boards are ordered and can hand out 3 for free and in europe I ship for free too. Please only ask if you really have time to work on it.

I have Tom Wiltshire ADSR and also figured out somewhat how the adsr's on the JP4 work. They have however slower min timings, but I think that can be adjusted to match the JP8 specs. The cap is charged with tiny pulses, the frequency of the pulses determines the timing, in essence really simple.

The Tom Wiltshire ADSR can of course be adapted fairly easy so I will personally start with those, I have them lying around. The board has the original IC footprint so anything needs to be adapted anyway. At some point when I have the time I will definitely try out the JP4 adsr also.

[ACME]

Tom Wiltshires ADSR works well, but I have decided against digitally generated envelopes because they don't correspondent the analogue character of the instrument.

[ACME]

wow that looks stunningly accurate, brilliant! Can I ask you some questions? I have a lot

Feel free to ask me, I also have a lot of questions.

Markus

Ok cool, can you verify my resistor analysis


There are 3 types on the board

- R25G, user groove suggested carbon 1/4 watt 2% On board pics I see red band at the location so I go with carbon 1/4 watt 2% as suggested.

- Carbon R25 , carbon 1/4 watt, gold band on pics > carbon 1/4 watt 5%

- Metal Film RB25FX, 1/4 watt, brown band > MF 1/4 watt 1%

Did you came to the same conclusion?

Do not worry about these resistors. Instead, just use 1% metal film.

Markus

[dikkietrom]

@ACME : I tend to agree, how are the 3310's working out for you? Were you able to check some differenes with an IR3?

Ironically Tom's pic could have an external cap and perfectly implement the JP4 design I guess. How analog is the JP4 adsr anyway one can wonder with the pulse resolution fed cap. I am pretty sure the IR3 implements this. The JP8 looks like a JP4 on steroids and ironically the JP4 sounds "better" to many. I don't know about that, My JP8 sounded great and I just want that back.

JP4 discrete ADSR, the question is how can we get decay time to 1ms instead of 14ms. Attack is actually 0.6ms vs 1ms

[ACME]

@ACME : I tend to agree, how are the 3310's working out for you? Were you able to check some differenes with an IR3?


JP4 discrete ADSR, the question is how can we get decay time to 1ms instead of 14ms. Attack is actually 0.6ms vs 1ms

By lowering the discharge resistor values?

The 3310 have some differences on the top end times. They must be compensated by an external circuit to match the R01. There are some nonlinearities in the middle, but the R01 shows this behavior too. Otherwise, the 3310 provides a clean envelope.

[KSS]

JP4 discrete ADSR, the question is how can we get decay time to 1ms instead of 14ms. Attack is actually 0.6ms vs 1ms

Look at the upper left corner of page 23 of JP-4 service notes. The differences between the A clock gens and D,R clock gens are your clues and cues.

[dikkietrom]

Do not worry about these resistors. Instead, just use 1% metal film.

Markus

Thanks good to know it does not matter much to you, I will go for 5% for the sake of it.

[dikkietrom]

The 3310 have some differences on the top end times. They must be compensated by an external circuit to match the R01. There are some nonlinearities in the middle, but the R01 shows this behavior too. Otherwise, the 3310 provides a clean envelope.

Can you eleborate the nonlinearities cause I don't understand since the env is by definition expo.

So you have a R01 to check against, wish I had that. Might buy that mks7 eventually, these things are impossible to find so far. I have slowly come to the conclussion that there is conceptually not really much difference between what happens in a JP4's env and Tom's pics. Both have a certain output resolution but the JP4's must be much greater. What's what? The pic is voltage controlled digitally generated and the JP4's is kind of digitally controlled analog generated. The 3310 also has logic, does that make it digital, do I care? I personally can't care less if it sounds genuine.

Here is the core concept, logic can be replaced with MC and timing clock can be replaced with OTA based voltage controlled resistance. I tried the original design but could not get it to work so I dumbed it down somewhat to understand what's going on exactly.

Edit : it is a link to an interactive scheme, click on it to see it. There is no other way to links to it from here.

https://lushprojects.com/circuitjs/circ ... Cshigh%5Cs[AND%5Csgate%5CsTrigger%5Cs%5Cp%5Csflip%5Csflop%5Cscounter]%5C%5Cn----%5Csopen%5Cstrigger%5Csswitch,%5Csdischarge%5Csto%5Cs0%5Csand%5Cstrigger%5Csswitch%5Cswill%5Csclose%5Csbecause%5Cstrigger%5Csswitch%5Csnow%5Cshas%5Csbeen%5Csopen%5Cssince%5Csgate%5Csis%5Cshigh%5C%5Cnif%5Csgate%5Csis%5Cshigh%5Csand%5Csvoltage%5Csnot%5Cstouched%5Cs5%5Cssince%5Cstrigger%5Cs[schmitt%5Cstrigger%5Cs%5Cp%5CsAND%5Csgate%5CsA]%5C%5Cn----%5Csopen%5Csattack%5Csswitch,%5Csdischarge%5Csto%5Cssustain%5Cslevel%5C%5Cnelse%5Csif%5Csgate%5Csis%5Cshigh%5Csand%5Csvoltage%5Cstouched%5Cs5%5Cssince%5Cstrigger%5Cs[schmitt%5Cstrigger%5Cs%5Cp%5CsAND%5Csgate%5CsD]%5C%5Cn----%5Csopen%5Csdecay%5Csswitch%5C%5Cnelse%5Csif%5Csgate%5Csis%5Cslow%5Cs[AND%5Csgate%5CsR]%5C%5Cn----%5Csopen%5Csrelease%5Csswitch%0Aw+672+208+672+256+0%0Ax+642+98+692+101+4+24+OUT%0Ao+43+4+0+4098+5+0.00009765625+0+2+43+3%0A38+1+0+1+10000+ADR%5CsTime%0A38+32+0+0+4.5+S%0A

[dikkietrom]

JP4 discrete ADSR, the question is how can we get decay time to 1ms instead of 14ms. Attack is actually 0.6ms vs 1ms

Look at the upper left corner of page 23 of JP-4 service notes. The differences between the A clock gens and D,R clock gens are your clues and cues.

Thanks, I see no page numbers but could find them on the module controller. There is indeed a 10k vs 4.7k diff between them. I also cannot get these simulated so I have no clue yet about the clock speed which would say something about the actuall difference between tom's pic and this approach.

[KSS]

Believe there is also cap differences. FWIW your last post is just a bunch of text in one block. What is it supposed to be?

[KSS]

Did see the lushprojects link is using the Falstad sim. You'll do yourself a favor to upgrade from that one to something like LTspice. Falstad is fun and can be useful ,but fails in ways that are not obvious and can be misleading even moreso than sims almost always are anyways.

And yes, the A clock gens use 150pF timing caps while the D,R clock gens use 47pF.