Because what the world really needs is another Baby 8 I’ve started on my own and would appreciate any constructive criticism as I begin to fumble my way through the synth world!
It’s a bit of a grab bag of elements from various places so I’ll cover what I’ve gone with and why:
The 555 implementation is working as is on a bread board and gives between 99% and 85% duty cycle as you go from slow to high pulse rates. This is used later to retrigger the gate. Whilst high 90s would be nice I doubt you’ll notice it as the tempo rises. The C anti log taper pot should give a nicer reponse than a linear. I tried “faking” one with a B100k and a 22k which works nicely but the ~20k result was a bit low. I’ll get a C50K to try as that seems to be the only easy to get C pot.
C11 is there as a hardware debounce on the switch. I need to test this still (need to find a switch), but I’ve used similar before.
I’ve added a regulator to knock the working voltage for the 555 and 4017 to either 9v or 5v as that’s still a decent range and make tuning a little easier.
I know I’ve not bypassed the TL074, if there’s room on the PCB I’ll add them.
I am not sure about the clock section. I’d like the output to link 2 Baby8s if I build them or act as a thru if I input the clock from something else. Do I need to more to defend here against odd voltages or to make sure it’s in spec (5v out?)
Gate is straight up stolen from above! Though I am not sure I actually need the Gate bus as if I’m not switching them it essentially becomes a 2nd clock out which would save 8 diodes, and the 2nd NPN.
I’d keep the transistor so you know what voltage level you’re dealing with regardless of what mangy old module you get the clock signal from. Then, as you say, drop it down to 5 V.
I’d also add a 1N4148: Cathode to Q1 pin 2, anode to ground. That will protect the transistor in the event that mangy old module turns out to be putting out -12 V.
Then I’d get rid of U4C and use the transistor output to clock the 4017 too.
I’ll explain what I think is happening as I want to learn if I am wrong:
R7 is there to limit current to the base
D20 will conduct if clock is -ve and stop weird things happening.
R16/R6 limit current through Q1 and also form a PD to tap off ~4.5v (4.3v measured with no opamp in a test and about 9.3V as V+ from a wall wart)
U1B then buffers that 4.5v out
R10 current limits the buffered out.
The voltage at Q1 E will 0v when clock is low and then V+ (ish as the transistor will have some resistance) when clock is high. This means I now have known voltage at Q1 E so I can use a simple PD to tap of a safe 4.5v for the Gate/Clock out which will also now feed into the 4017. Not sure if I should tap for the 4017 before the op amp though to ensure it’s not interfering with the outputs.
If R8 was 2k7 it would be much closer to 5V but is a value I don’t have and so it is an extra BOM item. 4.somethingV probably being close enough.
I think I’m going to need to re-annotate things as its “a bit of a mess” after all my changes.
Yes, do that. The signal from the voltage divider is fine to go to the 4017, it would not be so fine if you had it as above and the gate output were grounded (as it generally will be momentarily when patching) or otherwise funny. Basically when you have a current limiting resistor going to an output it should go only to the output, otherwise whatever else it connects to is exposed to the output jack.
Thanks, I’ve done that. And seeing as I want what is essentially 2 clock outs though once is intended to act as a gate I’ve buffered both in isolation (which means swapping back to a TL074).
Think this will work now. No it doesn’t as the clock is too low a Voltage to trigger the 4017. Please see: Understanding 4017 clock input
Yes. Any output should have a series resistor, 1k unless there’s good reason for another value.
A couple purely stylistic points about the schematic:
If you’re using global labels, they can be input or output shape. (Or bidirectional, tri-state, or passive, but I mostly ignore them.) So for instance the CLOCK coming off R7/R8 can be output shape while the CLOCK going into the 4017 can be input shape. It makes it easier to follow the signal flow. Same with hierarchical labels, if you’re doing multi-sheet schematics. Local labels just look like text and have no shape so can be harder to follow (than properly oriented global labels, but easier than non oriented ones) though sometimes I think they look better than global ones.
You can also help by making signal flow left to right when possible. For instance, I always orient my output jacks the other way, with the signal going in from the left. That makes it obvious at a glance which ones are inputs and which are outputs.
1k added but it’s made me realise the TL074 is in an incredibly daft place meaning some stuff is travelling the length of the board twice to and from a buffer. I’ll need to fix that, the trouble with auto routing I suppose as it hides some of this from you.
I didn’t even know the labels had direction. I’ve updated them all appropriately.
I’ve been learning Kicad rather adhoc googling stuff as and when I need to solve an issue. As such I’m probably no where close to “accepted normal” for a lot of things. I’ve been redoing all my custom library stuff to use going forwards as my old one was “a mess” honestly.
Hi Soundbender. I’m having a go at a layout of your schematic for eurorack - 10hp.
Couple of questions if that’s ok?
1st, from the schematic it looks like Diode 8 should go to pin 8 of the TL074? Is this correct or have I missed something?
The other thing is that you show on/off/on switches as an idea but from the photos it looks like you used two sets of on/on? Is that right, and if so was there a problem with on/off/on? I used some for a very basic 4 step as a test and they seemed to do the job…
I’m working with stuff I only partially understand here so be gentle with me…
for the switch yesI did not use on/off/on on step 1 & 2 (impossible for the 4017 to make a sequence of 0 step, and a sequence of 1 not sure that it is impossible but …)
therefore the sequencer can produce sequences of 2,3,4,5,6,7 and 8 steps
so for 1 & 2 on/on will do and on/off/on for the other 6
for the diode yes indeed there is an error I had to build it like that (and that’s work !?), I will modify that, thank you!
and sorry
EDIT : stripboard layout update on my blog & in the “verified stripboard” thread.
Will do. I hope to post an unverified layout here in a day or two, then try a build.
In the meantime one more thing: It looks like your layout has two 47k resistors between 4017 pin 14 and GND, one up at the IC and another one at the end of the long blue jumper. Is there a reason? I can only find one on your schematic.
This is a really good exercise for someone like me. I had looked at most of the same schematics you built this on and I’m learning lots from your build and this thread.
Honestly it’s quite difficult to get back into that, it puts my brain upside down but after checking yes it looks like I doubled it, surely only one 47k is enough (?), even if I built mine from the stripboard layout not the schematic and it works fine…
It annoys me a little to know that I made mistakes, but since it worked it seemed to me verified and I had shared it as such.
I really hope there isn’t another
again sorry for that and thanks for seeing this !
Yes, I figured that wasn’t going to stop it working. Thanks so much for replying, I’ll get on and finish my layout now and see if any of you kosmo obsessed lot are interested…
I had to build it like that (and that’s work !?), I will modify that, thank you!
but after checking yes it looks like I doubled it, surely only one 47k is enough (?), even if I built mine from the stripboard layout not the schematic and it works fine…