Look back in the thread, there's an excellent, labelled photo of the switch. If you look at the schematic, you'll see that each of the positions on the rotary switch is designated 0,2,4 . . .18,20. There's a contact on your switch for each. There should be two wafers, "a" and "b" on your switch, each possessing 11 contacts and 1 pole. the pole will be staggered in some fashion to set it apart from the switch contacts.
525 build thread
- Thread starter peter purpose
- Start date
Help Support GroupDIY Audio Forum:
peter purpose
Well-known member
Switch 5..wafer A.. position 20
thanks a lot for replying
i m sorry but there is one thing i don't understand
on the picture we can only see the following position for the switch:
fccw 0 2 4 6 .... but no sw5...
where is sw5 and which one is waffer a and which one is waffer b please?
because i dont see indication on the switch rotary
if you have a new picture , that would be helping
thank you guys
i m sorry but there is one thing i don't understand
on the picture we can only see the following position for the switch:
fccw 0 2 4 6 .... but no sw5...
where is sw5 and which one is waffer a and which one is waffer b please?
because i dont see indication on the switch rotary
if you have a new picture , that would be helping
thank you guys
peter purpose
Well-known member
SW5 is the Rotary switch designation in the schematic.
The wafers can be either way round, just know yourself which is which.
The wafers can be either way round, just know yourself which is which.
peter purpose
Well-known member
All connections from the switch go to the pcbs.
dmp
Well-known member
I'm just finishing building up a pair of these. Great project.
I've been messing around with the attack and switching the pot to 1Meg seems to work. I simulated this part of the sidechain and moved the 0.1uf cap (c3) between the attack pot and R5 and found that attack went from very fast to about 50 ms ,without noticably changing the release (the addition of 1meg will slow the discharge of C3, but the release resitances are large).This picture is from the simulation for a cv to the FET. The larger the attack pot, the slower C3 charges. A nice thing about this configuration is when the attack pot is fully ccw (fastest), the attack pot is fully shorted - like an original 525 with no attack mod.
Just listened this morning after calibrating, and I can tell the attack is slowing as the pot increases. I don't have a good enough ear to tell how much, but if I figure out a way to measure I'll report back.
Anyone know of a supplier for Omeg pots in the US? I think this would be good - ECO1MLIN
I've been messing around with the attack and switching the pot to 1Meg seems to work. I simulated this part of the sidechain and moved the 0.1uf cap (c3) between the attack pot and R5 and found that attack went from very fast to about 50 ms ,without noticably changing the release (the addition of 1meg will slow the discharge of C3, but the release resitances are large).This picture is from the simulation for a cv to the FET. The larger the attack pot, the slower C3 charges. A nice thing about this configuration is when the attack pot is fully ccw (fastest), the attack pot is fully shorted - like an original 525 with no attack mod.
Just listened this morning after calibrating, and I can tell the attack is slowing as the pot increases. I don't have a good enough ear to tell how much, but if I figure out a way to measure I'll report back.
Anyone know of a supplier for Omeg pots in the US? I think this would be good - ECO1MLIN
Attachments
peter purpose
Well-known member
Gimme an address dmp
dmp
Well-known member
I measured the attack times for a few positions. Full cw (~0 ohms), 100k, 500k, and full ccw (1Meg). The pot in this config allows quite a but of adjustment. The fast attack time is probably faster than shown here because I measured this using a 1kHz sine wave and calculated the rms - so the minimum observable attack time was on the order of what is shown here.
Attachments
dmp
Well-known member
dmp
Well-known member
dmp
Well-known member
Here's what I've learned trying to calibrate the LED meter for gain reduction.
If you set R55 as in the calibration notes, the full scale meter voltage goes from 0 to about 200 mv, shown in the orange curve below (which is the measured meter voltage for an increasing sweep of compression, like I posted yesterday). The LEDs don't light except for really high levels of compression, if at all (this was my experience anyway). This range however is appx. as expected to go to a 200 uA meter as used in the original 525 (generally these meters have a 1kohm resistance, yes?).
The LM3916 circuit, however, is set to work with a full scale range of 0v-1.2v, shown by the brown line (this is from the comparator resitances in the datasheet). The markers show when the LEDs turn on.
The blue line is what I got for the meter voltage for the same sweep of compression with R55 cranked up. This gets all the LEDs working (again, in my experience). The horizontal dashed lines show when each LED turns on as the gain reduction and meter voltage increase. The first five LEDs turn on before reaching 2 dB of compression, then the next five roughly indicate 3,4,6,8,~12 dB of compression.
For future designs using this chip, I'd suggest adding a few more components to set the voltage range for the chip (R2 in the datasheet for setting the high voltage reference, pin6, and a voltage divider to set the low voltage reference, pin 4, instead of tied to ground).
This would allow the meter LEDs to be more effective in showing gain reduction. The way I have it now, the first 5 LEDS all go on together. But with the extra resistors, you could calbrate the 1st LED to turn on at low GR and the 10th LED to turn on at high GR, and be more spaced out in-between. The slope and offset of the brown curve could be shifted to align better with the blue line. Ideally, you would want a comparator chip that matches the curve of the blue line but I don't think any of the LM391x chips would do this perfectly (or better than the lm3916).
If you set R55 as in the calibration notes, the full scale meter voltage goes from 0 to about 200 mv, shown in the orange curve below (which is the measured meter voltage for an increasing sweep of compression, like I posted yesterday). The LEDs don't light except for really high levels of compression, if at all (this was my experience anyway). This range however is appx. as expected to go to a 200 uA meter as used in the original 525 (generally these meters have a 1kohm resistance, yes?).
The LM3916 circuit, however, is set to work with a full scale range of 0v-1.2v, shown by the brown line (this is from the comparator resitances in the datasheet). The markers show when the LEDs turn on.
The blue line is what I got for the meter voltage for the same sweep of compression with R55 cranked up. This gets all the LEDs working (again, in my experience). The horizontal dashed lines show when each LED turns on as the gain reduction and meter voltage increase. The first five LEDs turn on before reaching 2 dB of compression, then the next five roughly indicate 3,4,6,8,~12 dB of compression.
For future designs using this chip, I'd suggest adding a few more components to set the voltage range for the chip (R2 in the datasheet for setting the high voltage reference, pin6, and a voltage divider to set the low voltage reference, pin 4, instead of tied to ground).
This would allow the meter LEDs to be more effective in showing gain reduction. The way I have it now, the first 5 LEDS all go on together. But with the extra resistors, you could calbrate the 1st LED to turn on at low GR and the 10th LED to turn on at high GR, and be more spaced out in-between. The slope and offset of the brown curve could be shifted to align better with the blue line. Ideally, you would want a comparator chip that matches the curve of the blue line but I don't think any of the LM391x chips would do this perfectly (or better than the lm3916).
Attachments
kazper
Well-known member
emrr said:
I had ran that number myself in my notes....
When I was looking into the stuff back on page 18 or so, I was shooting for attack times like a 1176 compressor. The 1176 specs call out 20-800us attack and from what I've seen on ther graphs it looks like the 1M pot gives a 400ms attack with the 0.1uf cap.
Edited for clerical error on the post...
dmp
Well-known member
Usually the specs for compressors say "attack as fast as" or something like that. I measured the attack for my DIY 1176 rev D (using the same method as for the 525) as 1ms to about 7ms. Bear in mind that the 525 is supposed to have frequency dependent attack/release and the graph just shows the response to a 1khz sine wave.
Attachments
benlindell
Well-known member
Just wondering what you're using to measure and graph attack times. Always wanted to measure those myself.
kazper
Well-known member
Specs for the 525 call for Attack time is as fast as 15u/ or Attack Time: 15 micro seconds
Looking at the specifications of the LA2A it shows a attack of 10 Milliseconds and a release of 0.5-5 seconds.
What was your final attack measurement at pot max on the 1Meg Pot? My guess was 400ms based on the first graph.
Realizing the original project modification design was having the pot adjusted to ground before. What does the provided 100K or 10 K pot give you for adjustment? I ask that because that was what was delivered to those who bought as kits. It may be nice to know what the difference attack ranges it makes if you change the pot's around as you have found to work.
If I get some time tomorrow I'll finish wiring mine now that I have all my tools found after the move.
Last are you checking this with a scope on the R5 and C3 junction?
Looking at the specifications of the LA2A it shows a attack of 10 Milliseconds and a release of 0.5-5 seconds.
What was your final attack measurement at pot max on the 1Meg Pot? My guess was 400ms based on the first graph.
Realizing the original project modification design was having the pot adjusted to ground before. What does the provided 100K or 10 K pot give you for adjustment? I ask that because that was what was delivered to those who bought as kits. It may be nice to know what the difference attack ranges it makes if you change the pot's around as you have found to work.
If I get some time tomorrow I'll finish wiring mine now that I have all my tools found after the move.
Last are you checking this with a scope on the R5 and C3 junction?
dmp
Well-known member
Ben-
I've been generating a signal to send to the 525 and acquiring the data from the output using National Instruments data acquisition cards. I wrote some test programs for this. (I work with this kind of thing in my real job...).
kazper-
If you put in the 100k pot tell us what you think. It is an easy mod - just move C3 off the main board and solder it onto the meter board below CM1, with one leg on GND and one on the inner ATT hookup. Both wires from the ATT pads go to the main board without change. Just a little mod on a fantastic project.
Looking at the graph, I agree it looks like the 1Meg slows attack to about 400ms & 100k about 20ms, but really someone with good ears will have to try it out - which isn't me.
I did measure the voltage at the R5 & C3 junction a little bit, but that is not what I am using for the attack measurement. Just the regular output signal.
I've been generating a signal to send to the 525 and acquiring the data from the output using National Instruments data acquisition cards. I wrote some test programs for this. (I work with this kind of thing in my real job...).
kazper-
If you put in the 100k pot tell us what you think. It is an easy mod - just move C3 off the main board and solder it onto the meter board below CM1, with one leg on GND and one on the inner ATT hookup. Both wires from the ATT pads go to the main board without change. Just a little mod on a fantastic project.
Looking at the graph, I agree it looks like the 1Meg slows attack to about 400ms & 100k about 20ms, but really someone with good ears will have to try it out - which isn't me
. I did measure the voltage at the R5 & C3 junction a little bit, but that is not what I am using for the attack measurement. Just the regular output signal.
