MUFF WIGGLER Forum Index
 FAQ & Terms of UseFAQ & Terms Of Use   Wiggler RadioMW Radio   Muff Wiggler TwitterTwitter   Support the site @ PatreonPatreon 
 SearchSearch   RegisterSign up   Log inLog in 
WIGGLING 'LITE' IN GUEST MODE

ba662 clone
MUFF WIGGLER Forum Index -> Music Tech DIY  
Author ba662 clone
guest
i just finished making a ba662 clone which is approximately the same size as the original (its a bit taller, will get pictures up soon). i tested it in my x0x, and it seems to work fine. i just wanted to check to see if there would be any interest in a production run of these. i dont know exact price, but it would definitely be less than ebay prices for an original.
Paradigm X
How well matched are they ? I know my SH09 uses four 'selected' BA662s, although ive no idea what parameters are 'selected'

Excellent initiative tho, many thanks, ill prob grab 4 to keep for spares.

Many thanks
Ben
dougt
There are also "A" and "B" grade originals... I have original BA662As I can sell for $35ea to wigglers by the way...
Altitude909
I'm in for sure. Be just as happy to have the board files actually
new voodoo
ive got a few but could original poster "guest/openmusiclabs" post a photo?
isnt OpenML who made the Mixtape Alpha? I got one of those..not as flexible as id hoped but im sure you were hampered by physical size restrictions of cassette tape emulation.
pm me w prices anyone who has either new or old, id be buying in a lot of, say, 10, 15, or 20..depending on price.

thanks for notifing us and good luck with your project/clone there!
guest
ok, here is a picture next to a ba6110 for comparison (my ba662 was at home).



altitude - i will post board files and such as soon as its all verified.

paradigm x - im using pmp4201 matched pairs, which have an absolute maximum offset of 2mV. im not sure how that compares to the original, as the datasheet seems to be lost in the sands of time. but, im guessing its pretty close, and hand selecting could easily generate A and B varieties.
Paradigm X
Im guessing they will be closer than the originals!
Not that i know much lol

Put me down for four provisionally then please, depending on cost.

Will they be covered out of interest, or raw pcb style like shown?

Cheers guest, legend as always smile

thumbs up
Luap
Neat! Any advantage for use in a x0xb0x over the 6110? Or is this really intended for other things?
So, whats next, cloned IR3109's!??
negativspace
I might be interested in a bunch. we're not worthy

(edit: in my terms a "bunch" would be several hundred.)
schmidtc
Nice work! I'd only want 1 or 2 for Roland spare parts -- haven't seen this chip fail yet, but nice to have just in case.
Peake
I've listened to a comparison of an XOX box using a 662 and one using a 6110, and I could hear that the 662 sounded better, to me. The 662A is used in the Roland 121 filter, and I'd love to have at least a dual in my setup.
aladan
I'm up for "some" (5? 10? 20?) but it does depend on price. As I see it DougT has set the bar that I think you need to get under...

Cheers,
A.
guest
ok, im going to make a small run, and see what people think of them. if all goes well, it will be a readily available thing from OML.
jgb
Definitely interested! My old x0xb0x might need one. smile
guest
Paradigm X wrote:

Will they be covered out of interest, or raw pcb style like shown?


sorry i didnt answer this earlier. im intending to leave them raw PCB style. this is actually not the best way to do it, as oils or moisture could possiblly build up, or the board could short to something else (these are both minor concerns). but, it allows the user to hack the board more easily, and if it needs to be covered, a bit of epoxy, silicone, or clear coat would easily do the job.

i was just thinking that i can also make 3080 clones, although im not sure i can fit them in a DIP8 format. maybe a double stack.
sonicwarrior
guest wrote:
see what people think of them.


No word on the price? If it turns out to be cheapish (less than $10) I'm interested, too. When it turns out to be less than $5 I would even buy some for my stock (you never know when you need such things).
guest
i wont know price until i get production figured out. everything will be ready to go in maybe a month from now.
Altitude909
hows the DC offset with this? When I was building x0xb0xes I would put a trimmer and a 10 meg pot next to the 6110 so I could trim out the VCA thud, might be cool to have that on the little board so one could do the same thing
minisystem
How about a faithful open source clone of the Roland IR3109 next. smile
guest
thats the second request for the IR3109. i think its just 4 VCAs and an exponential converter inside, so shouldnt be too hard to figure out. i think i have an IR3109 kicking around as well. but, i probably wont be able to make it a drop in replacement, as the circuit will be too large. any suggestions on form factor?
minisystem
The IR3109 and the BA662 are cloned in the 80017A replacement sold by analogue renaissance. That's a big SIP package.

I'd settle for a smaller SIP package for new designs. I get that it would be next to impossible to cram it all into a DIP compatible package.
Luap
guest wrote:
thats the second request for the IR3109. i think its just 4 VCAs and an exponential converter inside, so shouldnt be too hard to figure out. i think i have an IR3109 kicking around as well. but, i probably wont be able to make it a drop in replacement, as the circuit will be too large. any suggestions on form factor?


I believe you have an IR3109 Shruthi filter board? thumbs up
Actually I was half joking when I suggested cloning that chip. But I do think they could be popular and find plenty of homes.
It's subjective of course, but these things are in some of the best sounding filters out there.
From factor? Even if a bit larger, I expect if it were somehow vertical, much like your BA662 clone, then it would still fit onto most boards that can use the originals.

Anyway, sorry for going off topic! I'd gladly try out one of your BA662 clones in one of my x0xb0xes. I have a couple of genuine BA662's I can compare with, although my testing wouldn't be the most scientific..
guest
Quote:
I believe you have an IR3109 Shruthi filter board?


yes, very true, but im a bit reluctant to take the shruthi apart! it sounds great in there. thanks again.
Jackdamery
guest wrote:
thats the second request for the IR3109. i think its just 4 VCAs and an exponential converter inside, so shouldnt be too hard to figure out. i think i have an IR3109 kicking around as well. but, i probably wont be able to make it a drop in replacement, as the circuit will be too large. any suggestions on form factor?


Two SMD LM13700s and a SMD TL071. You could probably get it pretty small. There are four resistors, one between the noninverting input of each gain cell and ground. I was trying to do this a while ago and with some help I measured the internal resistors to be 432 ohms.

Someone pointed out to me it's basically the same filter as the Roland system 100 but condensed into a chip. Hence FET buffers would probably do between gain cells. You can see this configuration if you look at the JP8 high pass filter.

For the expo converter I used Thomas Henry's standard current sucking circuit.

Here's a recording I made, when I built half of it, i.e. 12db attenuation.
Demo of 12db OTA IR3109
CJ Miller
Nice work!

guest wrote:
i was just thinking that i can also make 3080 clones, although im not sure i can fit them in a DIP8 format. maybe a double stack.


The 13600 seems close enough to the 3080, and is still available. But I think there will be interest if you can make an equivalent CA3280.

Money is too tight for me to buy BAs, since I don't really need any. But I can't afford to not buy some CA3280s in the future, since I have several projects on hold waiting for a reliable source. And I know I am not the only one.
Dimitree
3080:

http://www.diystompboxes.com/smfforum/index.php?topic=31109.0
IImyment
Hi Guest . I'll be up for three or four depending on the price...
isak
hi guys

i see the old x0x gang is here smile
cheers guest, you always surprise us!
limpmeat
I'm interested in a couple!
clusterchord
if it sounds dead identical like Roland i am interested in a dozen.



as for 3109 or anything else, any attempt at miniaturisation changes sound, as we witnessed with Rolands own designs. the juno106 is worlds apart from the sound of juno60 or JP8 filter. and its all supposed to be repackaged reminiaturised iteration of ba662 in cascade..


if youre cloning you should stick to same form factor, trace widths and everything else as close to original as possible. otherwise the idea, many of us had im sure, to use these as spares when something dies in our old Rolands , wont work. you would have a different sounding voice out of 8 etc.

just thinking out loud.


in any case insteresting developments, keep up the good work thumbs up
minisystem
guest wrote:
i wont know price until i get production figured out. everything will be ready to go in maybe a month from now.


Status update? smile
guest
guest wrote:
i wont know price until i get production figured out. everything will be ready to go in maybe a month from now.


well, that was a long "month". units should be coming back from production this week, and then i need to QA them and get them to the store. i should know pricing next week.
andybee
very nice!
please let us know, where they are avaialable!
thanks!

&

we're not worthy
4teenth
Ooooh....

I came here from the Yocto forums - I've got a poorly Juno 6 that needs a couple of these..

Any idea of a really rough ballpark figure, as I know someone that was selling a bunch of 662s and now only has a few left, and a pair is going to cost me about $80 plus shipping..

Worth holding out for your repro would you say?
dougt
4teenth wrote:
Ooooh....

I came here from the Yocto forums - I've got a poorly Juno 6 that needs a couple of these..

Any idea of a really rough ballpark figure, as I know someone that was selling a bunch of 662s and now only has a few left, and a pair is going to cost me about $80 plus shipping..


I have them in stock and can do $70 for 2 BA662As...
guest
4teenth wrote:

Any idea of a really rough ballpark figure


it will probably be in the 10$ - 15$ range.
4teenth
OOOOOH! Put me down for a couple then smile
guest
ok, so we finally got the ba662 clone in stock:
http://www.openmusiclabs.com/store1/store-ba662-clone/
Luap
Ooo, Might order me a couple tomorrow..
(Does it say "BA662 Clown" on the back though?? hihi )
guest
Luap wrote:
Ooo, Might order me a couple tomorrow..
(Does it say "BA662 Clown" on the back though??)


ummm, maybe. i remember seeing something similar on the x0x forums, i think from momerath, who had cereal instead of serial. i thought it was pretty funny and figured id incorporate such things when possible.
minisystem
Amazing! Nice work. thumbs up
4teenth
SWEET!

Just ordered a couple smile
4teenth
Mine have just arrived in the uk 8_)
4teenth
I'm happy to say that these work perfectly in a Juno 6 w00t

http://www.4teenth.co.uk/2014/04/fixing-missing-voices-on-a-roland-jun o-6/

[video]Yw4rHsIetRA[/video]

(How do I embed youtube vids?)
guest
awesome, thats great to hear!
delayed
guest wrote:
ok, here is a picture next to a ba6110 for comparison (my ba662 was at home).



altitude - i will post board files and such as soon as its all verified.

paradigm x - im using pmp4201 matched pairs, which have an absolute maximum offset of 2mV. im not sure how that compares to the original, as the datasheet seems to be lost in the sands of time. but, im guessing its pretty close, and hand selecting could easily generate A and B varieties.


Did you ever post board files anywhere?
guest
yes, all information is up on the wiki
http://wiki.openmusiclabs.com/wiki/BA662
delayed
nice write up. thanks for posting all this information. it is a good read.
Altitude909
So I've used these in x0xb0xen a few times now and was wondering how feasible putting an offset trimmer right on the board would be? Could I simply make the board taller and slap a 3mm right angle trimmer and resistor above the existing circuit?
guest
yes, thats totally fine, and would be a great improvement. if you really wanted to save space, you might even be able to squeeze it onto the back of the board.
Altitude909
That was my plan, make the board a bit taller and just slap it on the back
guest
also, i came across this a few days ago:
https://www.facebook.com/pages/BA662-clone/826416797440804?fref=nf

any guesses? im bettting a current mirror and replica output stage under the 16p IC (LM13700?). my reasons for saying that are: 1. the IC doesnt look fully seated, so there is room under it for SMT components. 2. if a special IC was fabbed, why use a 16P DIP casing? even if you couldnt get the original 9P SIP, a 14P DIP would be the next size up. or do a flip-chip or chip-on-board.
Altitude909
guest wrote:
also, i came across this a few days ago:
https://www.facebook.com/pages/BA662-clone/826416797440804?fref=nf

any guesses? im bettting a current mirror and replica output stage under the 16p IC (LM13700?). my reasons for saying that are: 1. the IC doesnt look fully seated, so there is room under it for SMT components. 2. if a special IC was fabbed, why use a 16P DIP casing? even if you couldnt get the original 9P SIP, a 14P DIP would be the next size up. or do a flip-chip or chip-on-board.


I'm scratching my head about that thing. A) It costs more than the real part, you can get NOS dotted A parts for less than 45 euro from several places B) giganto DIP TH part to SIP header? C) though hole, 1/4 watt resistor if there are other SMD parts underneath?
guest
yeah, im not sure what the resistor is about.

the price seems to reflect the time they spend to bin them all.
Altitude909
as it turns out, the trimmer I wanted to use was the first hit in the eagle lib so this was a snap, quick check?

You mention in the wiki that the offset bias could be trimmed out as well, there is plenty of room for another trimmer so any tips on how to do that as well?

* Edit fixed



guest
the 10M should go to pin2 or pin3 to trim the input offset voltage (usually pin2). to trim out the output offset voltage, reduce R1 to 10k, and put a 100k in series with it. then you can vary the current through the transistor to set the Vbe to match. this doesnt matter a whole lot for the x0x, as the ouput is AC coupled.
Altitude909
grr. i cant even follow my own damn notes. Will revise.
Altitude909
fixed for pin 2. I'm not going to mess with the output offset. It would require some layout changes and I am far from comfortable moving stuff around on something this dense
Jackdamery
Anyone got an eagle library part for the BA662?
Altitude909
Just use a SIL7 header, it's standard 0.100" spaced pins
Altitude909
And here is my modified one with the offset trimmer onboard:

Guest: With your OK, i'd like to share this iteration as an OSHpark project..



guest
awesome work! let us know how the trimmer works.

and sure, feel free to share the files
Altitude909
https://oshpark.com/shared_projects/QCnNP4oK

Thankfully my x0xb0x has all the parts socketed around the VCA from testing this and that way back when so it shouldnt be too much of a chore to get this tested
Jackdamery
I just received my PCBs from OSH Park, does anyone in the UK want a couple?
ablearcher
So, my juno 6 had a dead ba662 in one of the filters and instead of buying guests clone i decided to pick up a few of altitudes version from osh park and try building them... by far the tiniest pitch smd i have worked on.

So when i put mine in all the lights in the oscillator section go off and no sound comes out. Im guessing i have a short somewhere even though all the joints look okay.

Anyone have any luck building these and/or using them in a juno 6?
guest
most likely a short or flipped transistor/wrong transistor. it is really small SMD. you might want to probe around to see if things are connected that shouldnt be. or just rehit all your joints with some wick and try it again.
Altitude909
I've built about 30, it's tedious work and you really need a very good way to inspect the parts. SOT parts are prettty challenging to do by hand
synthdude
Out of the four that I have built so far, only one of them works. I can't imagine making 30 of these, thats impressive if you are doing them by hand! I think the best way to handle these is to have them made instead of trying to make them yourself. This should not be too expensive.
guest
i have to imagine its cheaper to just buy them, rather than have them made (for small quantities). we recently dropped the price to 9$. granted, they dont have the trimmer on them.
Altitude909
Yeah, wont be doing any quantities by hand anymore. It was a one time thing for RE-303 alpha ppl.
jhulk
HOW ABOUT THIS TH VERSION IT USES T0-71 SIZE METAL CANS OR YOU CAN MATCH

2n3904/2n3906 for cheap

easier to solder 40mm/20mm
jhulk
i did these for the jup4 modular voice im working on
ringroad1
jhulk wrote:
i did these for the jup4 modular voice im working on


ooi did you test 3080 in for the 662?
guest
looks great. is the bias trimmer in there for output offset voltage nulling?
guest
jhulk wrote:
i did these for the jup4 modular voice im working on


i think you can use an LM13700 for most of the JP4 voicing. the only difference between the core of the LM13700 and BA662 is the control input. but, if this is driven with a current source, then that doesnt matter. the buffers arent used in a lot of cases in the JP4 as well, and FETs are used instead.
jhulk
the ba662 is very similar to the 3080 current runs a little different

but are similar the buffer on the ba is different to the lm13600/700 as its a npn/pnp type on the ba and its a wilson current mirror on the lm

i fitted both trimmers one for input bias and one for output bias on the th version

and your right the buffers on the jup4 have no connection so i could do away with those connections just need ota cell

there are a few diffeent types of ota core emu used a few different ones

the ssm2030 the current mirror in box is a ota its based on the ssm2020 cell
which is a diff current pair connected to the ota core the buchla 158/258 discrete versions are very similar in discrete designs to the emu design

the same design is also used in the ssm2040 filter core theirs a an data sheet for the ssm2020 and it has circuits for filters vca and oscillators vc eq rm and am modulators.

all using the same type of ota cell it was very similar up unto the ssm2124 they changed the design when they went to the ssm2164

i have done all of these as a discrete component version.

i also have a version of the ca3280 i would like you to look at guest as its a very involved chip with lots going on
guest
jhulk wrote:

i also have a version of the ca3280 i would like you to look at guest as its a very involved chip with lots going on


cool, would love to check it out. my concern with the CA3280 is that i felt it needed matched quads to really do well. also, you can get almost identical performance from an LM13700 plus an opamp.
loki
Jamie,

There is not really any point in using the LS318 and LS358 in your BA662 clone. Log conformance isn't an important characteristic of transistors in a current mirror. For transistors of the same geometry, delta Vbe and beta to a lesser extent are everything you want to know.

For matched transistor pairs you might try the MMDT4401 npn pair and the MMDT4403 pnp pair. These transistors have larger base-emitter area and, as a consequence, low rbb and low rbe. The rbb has a significant effect on the noise performance of the transistors. The late Marshall Leach posted material about low noise audio design on his web site. Still available. An editor at an audio magazine he did a preamp design article for recommended the 2N4401 and 2N4403 as a secret ingredient of the mixing console designers.

Reference the National Semiconductor Discrete Databook from 1978, available at the Internet Archive, for die photos and characteristics of different transistor processes.
ablearcher
After reworking the BA662 clone a couple times I gave up and ordered it assembled from from synthcube.

Same result, inserting it into the juno 6 (IC18) causes the oscillator leds to turn off and no sound will come out. Remove it and the juno 6 works fine except on some of the voices the filter will not resonate correctly.

What's going on here? Does the clone not work in this case or am I missing something?
guest
i will take a look at the juno6 schematics and see if there is anything that might be wrong. all of the synthcube stock is tested before shipping, so it should have been a good unit. a couple of quick ideas:

1. is orientation correct?
2. maybe something else is blown in the circuit that caused the original ba662 to get blown?
guest
my schematics dont have good labels. is it the resonance feedback control amplifier? the one with R2 and R3 connected to its inputs?
guest
assuming its the resonance control amp (which seems very likely), then there shouldnt be any problems using it there. with the amp removed, measure the voltage on all of the pins in the socket/board. they should be:

1. 0V most likely (its current, not voltage at this node)
2. 0V with 50mV signal
3. 0V with 50mV signal
4. -15V
5. -15V
6. 0V with 50mV signal
7. +15V
8. +15V (maybe a bit less)
9. +15V

its most likely TR1 blown. you can test this node without inserting an IC by putting a diode between pin1 and pin5 (anode to pin1), and seeing if it does the same thing where the leds go off. you can also just put your meter in current mode (start with a high current setting - 1A or more) and measure the current between those pins. if its more than 1mA there is a problem.
dougt
ablearcher wrote:
After reworking the BA662 clone a couple times I gave up and ordered it assembled from from synthcube.

Same result, inserting it into the juno 6 (IC18) causes the oscillator leds to turn off and no sound will come out. Remove it and the juno 6 works fine except on some of the voices the filter will not resonate correctly.

What's going on here? Does the clone not work in this case or am I missing something?


I still have original Roland BA662s in stock if you want to get one.
aabbcc
jhulk wrote:
HOW ABOUT THIS TH VERSION IT USES T0-71 SIZE METAL CANS OR YOU CAN MATCH

2n3904/2n3906 for cheap

easier to solder 40mm/20mm


Did anyone try this with 2n3904 and 3906? I know synthcube sells the ba662 clone but this looks like a fun little project smile
tupinamba
jhulk Where could I get a few of those BA662 clone boards? I like the non-smd approach... ;-)
Thanks!
Jaytee
Pretty sure jhulk stopped posting here a while ago. Try the Logan Soloman Facebook group.
tupinamba
Ok, great, thanks for the tip! So jhulk is Jim?...
Cheers!
Jaytee
I think jhulk is Jammie/Emma.
flts
Jim Soloman is/was aquatarkus on this forum (I think?), he's the other person in the duo. Emma Logan is jhulk - or actually, Jammie Logan is jhulk, but he's posting from his wife's account.

Or that's how I've understood it at least...
aabbcc
flts wrote:
Jammie Logan is jhulk, but he's posting from his wife's account.


And I don't even know where to start about how weird/creppy/Idon'tknow/strange that is seriously, i just don't get it

Anyhow those BA662 boards would be very easy to recreate and order and would good first eagle/kicad for anyone smile
tupinamba
OK, ha ha, this is fun!
Yes, those boards are a pretty simple Eagle project, but if the Eagle or Gerber files already exist, or if Jammie has some spares... :-)
Maffez
guest wrote:
jhulk wrote:
i did these for the jup4 modular voice im working on


the only difference between the core of the LM13700 and BA662 is the control input. but, if this is driven with a current source, then that doesnt matter.


hey ho, sorry for necroposting and asking a potentially dumb question

by difference you mean pin 4 of BA662? thinking about an lm potentially utilized as vca in an sh101 circuit

cheers!
LED-man
Just remember the pcb won’t fit in all circuits because of the size.
guest
the LM13700 has the almost the same core, but a completely diffferent buffer (pins4,7,8), so the clone is really the best bet for the 101 (been used in plenty at this point).

as for the size, it is slightly taller due to the headers used as pins. but, if you just used bare pins, it would be the same size:

MUFF WIGGLER Forum Index -> Music Tech DIY  
Page 1 of 4
Powered by phpBB © phpBB Group