I'm trying to modify an analog delay effect that is used in a pinball machine to be able to adjust the delay time up or down somewhat. It currently has a fixed delay time. I have already modified the circuit to add broader control of the feedback and put in a trim pot to control the output mix. Here's the schematic...
I don't really understand how the clock circuit works enough to know where I might look to adjusting the delay time. Can anyone help with a technical explanation and maybe a recommendation?
Also, there is some very audible clock noise bleeding into the audio circuit that I can clearly see on my scope. I assume this would be due to poor component layout as it seems to be problematic on other games that use this effect. Any hope of fixing this?
I don't really understand how the clock circuit works enough to know where I might look to adjusting the delay time. Can anyone help with a technical explanation and maybe a recommendation?
Also, there is some very audible clock noise bleeding into the audio circuit that I can clearly see on my scope. I assume this would be due to poor component layout as it seems to be problematic on other games that use this effect. Any hope of fixing this?
Have you considered monkeying around with the DC control voltage applied to the phase locked loop's (chip U5) VCO-input pin (pin 9)?
It appears that U5 generates a square wave timing signal, then quad-NOR2 chips U6 and U7 turn that into a non-overlapping pair of "two phase" clocks (TP7 and TP8). These clocks are applied to the delay line IC.
You could connect a dual channel oscilloscope to TP7 and TP8, to measure the period and the nonoverlap (dead) time. Snap many many scope photos for later analysis at your leisure.
It appears that U5 generates a square wave timing signal, then quad-NOR2 chips U6 and U7 turn that into a non-overlapping pair of "two phase" clocks (TP7 and TP8). These clocks are applied to the delay line IC.
You could connect a dual channel oscilloscope to TP7 and TP8, to measure the period and the nonoverlap (dead) time. Snap many many scope photos for later analysis at your leisure.
Having worked in that industry many years, I have to caution you. It is just a pinball machine, not a recording device. With all the clanking and popping and thumping going on, the last thing they were concerned with was a little extra noise on the echo circuit.
Are you using this not in the pinball? If so, did you lump all the grounds together? In the system, the digital and analog circuits use separate grounds. One reason for this is noise abatement.
Agree with Mark, you need to change the clock speed down at U5.
Are you using this not in the pinball? If so, did you lump all the grounds together? In the system, the digital and analog circuits use separate grounds. One reason for this is noise abatement.
Agree with Mark, you need to change the clock speed down at U5.
Thanks for the replies!
I've definitely not lumped any grounds together. I haven't done anything to alter the factory layout, just replaced some electrolytic caps, a resistor and added a trim pot. The clock whine (seems to be about 10khz and very slightly oscillating) was there before and is still there now. If it can't be dealt with, it will be fine in the bar where it will eventually live. If there's a chance that it could be helped however, I see no reason not to try.
I've definitely not lumped any grounds together. I haven't done anything to alter the factory layout, just replaced some electrolytic caps, a resistor and added a trim pot. The clock whine (seems to be about 10khz and very slightly oscillating) was there before and is still there now. If it can't be dealt with, it will be fine in the bar where it will eventually live. If there's a chance that it could be helped however, I see no reason not to try.
What power supply circuit are you using to power it?
You might - or might not - find this interesting. MAny years ago, my friend had a Gottlieb Funland mechanical pinball. (Remember the old WIlliams baseball pinballs? You had a bat down at the bottom, and a big button, a pitched ball came up from under a flap and you batted it towards the field?) I took the sound unit from a Williams baseball game, and stuck it in his Funland. I had to add some switch blades to the relay stacks, but I managed to wire it to the game functions. Higher score targets got explosion sounds. Getting the specials lit turned on the background cheering, and winning a replay got the loud cheers and explosions. His girlfriend let me in while he was away so I could do this. He came home, went down to play some pinballs, and all of a sudden his game was cheering and making explosions.
You might - or might not - find this interesting. MAny years ago, my friend had a Gottlieb Funland mechanical pinball. (Remember the old WIlliams baseball pinballs? You had a bat down at the bottom, and a big button, a pitched ball came up from under a flap and you batted it towards the field?) I took the sound unit from a Williams baseball game, and stuck it in his Funland. I had to add some switch blades to the relay stacks, but I managed to wire it to the game functions. Higher score targets got explosion sounds. Getting the specials lit turned on the background cheering, and winning a replay got the loud cheers and explosions. His girlfriend let me in while he was away so I could do this. He came home, went down to play some pinballs, and all of a sudden his game was cheering and making explosions.
Those baseball games are called Pitch and Bats. I have friends that collect them. They're a lot of fun!
The power is supplied to the Say it Again board via the 43v solenoid bus at which point it is regulated down the 12v to power the circuit. It is powered separately (using a different power bus) from the Squawk & Talk sound board to which the reverb is connected. The analog grounds are connected between the S&T and Say it Again boards at the audio IN/OUT connector J2 and the digital grounds are connected between the two boards at the power connector J1.
Full schematics for Centaur are here in case someone wants to look at more... Internet Pinball Machine Database: Bally 'Centaur'
This works... I changed R39 to a 250k pot and get a good variation on delay time from fairly short to longer. The good news is that high pitched noise from the clock circuit gets higher and less audible as the delay times get shorter but the bad news is it becomes very loud and much more audible as the delay times get longer. I prefer the sound with the shorter delay time (what I was looking to achieve going in) so this is good for me but for other people with original boards, it would be nice if there was some solution to filter or eliminate the noise from the audio.
SAD4096 was a VERY common BBD delay line series. I worked with the first run (a 256-sample line) and while interesting it made bad audio. While clock whine can be worse with bad layout, they whine no matter what, and at long delays it is in the audio band.
BTW: a known-genuine SAD4096 from 1981 may be *very* valuable on its own. The unique sound spawned a type of guitar pedals which just will NOT go out of style even as FAR better delays became available. Original stocks are long gone, eBay is full of fake "SAD4096". CoolAudio now makes a 4096 class BBD chip so that Berhenger can clone the old classic pedals, but purists want the vintage chips.
There's been much effort to get better sound out of the flawed BBD concept. Pre/De-emphasis and companders, which your build lacks. However rather than re-hammering an old wheel, it may be best to shop the rock-shop for a delay pedal.
BTW: a known-genuine SAD4096 from 1981 may be *very* valuable on its own. The unique sound spawned a type of guitar pedals which just will NOT go out of style even as FAR better delays became available. Original stocks are long gone, eBay is full of fake "SAD4096". CoolAudio now makes a 4096 class BBD chip so that Berhenger can clone the old classic pedals, but purists want the vintage chips.
There's been much effort to get better sound out of the flawed BBD concept. Pre/De-emphasis and companders, which your build lacks. However rather than re-hammering an old wheel, it may be best to shop the rock-shop for a delay pedal.
The delay time is proportional to the clock period.
In case of SAD4096 with 4096 "buckets" a 50us clock period yield 50x4096us=200ms delay.
50us clock period equals 20khz clock frequency
So if you expand delay to 400ms you require 100us clock period corresponding 10khz clock frequency.
That is what you hear with increasing delay time.
All in all, these analogue bucket brigade delay lines have always been a poor solution for longer delays. There is not much you can do about that except cutting high frequencies using a low pass.
In case of SAD4096 with 4096 "buckets" a 50us clock period yield 50x4096us=200ms delay.
50us clock period equals 20khz clock frequency
So if you expand delay to 400ms you require 100us clock period corresponding 10khz clock frequency.
That is what you hear with increasing delay time.
All in all, these analogue bucket brigade delay lines have always been a poor solution for longer delays. There is not much you can do about that except cutting high frequencies using a low pass.
I attempted to do this but my implementation was incorrect... put a basic resistor cap network at the output and the volume dropped dramatically and it seemed as though my 6khz filter (based on the calculator) was closer to 600 hz. I'd love to try this mod so if you have any suggestions for ways to implement this I'm all ears.
In case you add the echo to the dry signal as an effekt, there is no point in designing the echo path with full audio bandwidth. You may cut high frequencies above 2..3khz.
This is not true if you want to delay the full audio signal, but in that case you end at the limitations of that antique BBD technique.
You can build a cheaper and even better analog delay using PT2399.
I know nothing at all about pinball machines, but if all you need is a stronger signal to feed to a guitar delay pedal, why not build a simple transistor gain stage to insert ahead of the pedal? (And a pot on the output to bring the signal level back down.)
Since this isn't exactly a Hi-Fi situation, and you don't have to have high imput impedance, a single transistor, three resistors, and a couple of small capacitors will probably do all you need. That seems a lot simpler than trying to re-invent an analogue delay using a chip as rare as a pregnant Dodo.
-Gnobuddy
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.