waggled my clocks

[duck_arse]

here is the boss CE1, with its 2 phase modulated clock. clik for biggger.


just for fun (thanks, Cozy), here is my bredborded translation.


(and my black and white bb translation) which now runs on +9V, gives a 50.5kHz, 9V peak, square wave (when it starts ....). attached to test is a simple single osc, you all know the one. and it works.

for a while. the modulation seen gets lesser as it goes longer, but with no visible voltage changes. why would this be? what have I not done right in the swaps.

and, on operation: at the D1//C2 (D2//C1) points, I get 8V p-p, with a +8V offset, and at the K's of D3, D4 and D5 I get 16V with the modulation voltage on top again! why is that? how's it work, and is it working right?

[anotherjim]

I think you have a voltage doubler/charge pump thru C9. It gets stored on C6 and has nowhere to discharge - the diodes are blocking it in.
Looks to me like the original is DC coupled and has no equivalent to your C9. My brain is wondering if you can lose that C9. Hmmm...?

[duck_arse]

thanks jim, I'll let you know after a big sleep .....

[duck_arse]

well, sleep had, but no insights gained. spent the day playing with varicap-ing diodes and cmos. the dc-ing aspect of this is beyond my pay scale, frankly. what are all those diodes for? the 16V appears at D5//R9 even without the mod-osc connected, so no C9 to charge.

[duck_arse]

now with dc coupling. I been playing at this today, and it works. should it? it is shown in its most cut-down state, the 4093 would be a 40106 for the extra buffering of the |Q/Q outs.

so now I ask: what does a chorus chip like the mn3207 need by way of clock-f shift to do chorusing good? this whole excercise is to find a non 4046/mn3101 solution to bbd drive. I'd prefer to get that load of transistors working properly than to reinvent some square wheels.

[edit :] it seems I forgot to add this .....

[anotherjim]

Note G on the CE-1 sheet says clock lives between 5 & 16us periods. Or 200kHz to 62.5Khz if I've worked the reciprocals right. 200kHz is oft quoted max for BBD's.
That should be the chorus range?
1024 stages x5us = 5(ish)ms to x16us = 16(and a bit)ms.
Sounds about right?

[ElectricDruid]

Note G on the CE-1 sheet says clock lives between 5 & 16us periods. Or 200kHz to 62.5Khz if I've worked the reciprocals right. 200kHz is oft quoted max for BBD's.
That should be the chorus range?
1024 stages x5us = 5(ish)ms to x16us = 16(and a bit)ms.
Sounds about right?

Yeah, sounds about right to me. 5ms to 16ms is definitely chorus.

200KHz is the datasheet maximum for the MN32xx BBDs, but in practice they'll go to double that without much trouble, and the output level only starts to drop off above 500KHz or so.

HTH,
Tom

[duck_arse]

so, OK, it seems the answer to the mess of transistors was to add another, between the nasty 16V on D5 and ground, and then jamb some lfo into its base. works "better", needs fiddle.

and it seems I had the invertor layout in the mn310x backwards, mighta been the wrong in my first attempt.

[anotherjim]

I guess that in the BJT astable that C1 & C2 are capable of pumping charge into C6 - unless something can clamp or bleed the charge off. Adding a BJT control would do it - as you have.

Q1 & Q2 are added to the basic astable to provide a current controlled variable frequency. D4 & D3 are stopping each phase hitting the other and combining the controlling currents in one node.

Not sure about D5. It might only be there in the Boss product because the control can go negative during power up/down. You might not need it with a single supply.

[duck_arse]

a shame, but the mess of transistors has now gorn. replaced with cmos, no schmitt! and it was working better/bigger with a wrong (10x) value.

[anotherjim]

Yeh, I was wondering if the transistor scheme has any advantage over CMOS gate drive to the BBD clock lines. It's a hard pull down with the Tr's but a weak pull up. CMOS is a fairly equal drive either way. So is there any advantage, especially pushing a BBD above 200kHz when the mass of gate capacitance starts to fight the clock?

[duck_arse]

stay tuned, jim, there are questions on CCO's coming soon (I found a different circuit).

[duck_arse]

todays variant is from the boss dd2, and uses a gate/inverter to replace the usual oscillating npn. it needs slower values, but it works chuffa-chuffa-chuffa on the breadboard. only it was doing strange things with the emitter follower voltage controling.

http://www.hobby-hour.com/electronics/s/schematics/dd2-digital-delay-schematic.gif
-and-
http://www.hobby-hour.com/electronics/s/schematics/boss-bf2-flanger-schematic.gif

but! the yamaha fl10 and the boss bf2 uses a pnp osc transistor, and two gates, with a current control arrangement to modulate the osc. which is nice. and works, also chuffa-chuffa, like a steam train, on the bb.



[yes, it's wide, it's also colour, sorry.] all well and good, but I'd like to use the 3 inverter from the boss, and add the current control of the yamaha, doing away with that oscillating transistor. but it won't chuff, just sits there spewing out a train of spikes.

is there a way to do the current control on the boss arrangement?

[duck_arse]

- bump -

any ideas? trying to replace the Q103 with an inverter? anyone?

[duck_arse]

at ease, people, we have an answer. reffering to the RHS of the rev00.00 above, the addition of a resistor, lets call it "R2", to V+ from Q101 collector, and a diode, lets call it "D2", in series with R104, K to D101 K, did the tricks. that circuit did work, would work for a flanger probably, but has so many tunables that it was unmanagable.

and every time I connected the second mod-osc, all the timings shifted with the DC offset. so, another one in the bin! I finally (as of yesterday) decided on a two (schmitt) inverter, one transistor osc, driven by an lm324 lfo/mixer, and all the extra inverters point at the BBD clock inputs.

[points awarded to anyone who can "see what I did there", where I did it.]

[Kipper4]

What you did?
You waggled the Arse off of it.
Do I get a sausage?

[duck_arse]

no sausage, no cigar.

use the force on the added components, it will become clear.

[duck_arse]

todays problems. shown is the circuit as it stands today. yesterday it had different values, and was swinging chuffa chuffa to specs. today I decided to fixxe some mixing- and divider-resistor values, and to reset V/2 for symmetric triangle. which I did, chuffa chuffa chuffa, only with better resistor values....



.... only now I see there is a cap on the bb that is not connected. so I did, and there goes all my good clock-swing. ...BLAST... I've just checked the original diagrams, and I now see that my problem cap, C1, connects to the base of the transistor I've replaced w/ the opamp. (the "ELUR K-IV CHORUS" does have that cap on the outputting divider, but it's the only one I can find at the moment.)

so now my question is, given the carp that flushes back from the osc to the opamp without that wrong cap, should it stay or should it go, now? and if I stay, what value?

[anotherjim]

C1 looks in the right place. Value depends how fast the LFO runs. Too big though, and I'm sure you see it, is the Triangle isn't a triangle if you cut it's harmonics. But at fast LFO, can you hear difference between tri and sine?

LFO pass cut-off is R5/C1 while carp swimback is R101/C1. The bigger C1 the better I would think for keeping those pesky fish out of duckpond. Probably start with C1 value as theoretical cut off of R5/C1 at max LFO freq' then reduce it a value step at a time.
BTW, Wot you did above was make it like the other one. Easy.

[duck_arse]

tri and sine - erm, I think there is a difference, but I don't know if it's just from all the unbalances that I've had come round and play on the bredbord, or real physics. or carp. mostly I watch the f/f output to bbd, and as soon as I have the full documented swing, then add the 10nF, it reduces to not much at all swing. possibly with either waveshape, I'd have to write notes on that next the cro is on.

and the 10nF was the correckt value at the end of a 220k resistor.

like the other one, but subtley different.