Harpo

Re: sontec 1 RU build thread
« Reply #760 on: January 31, 2019, 05:16:55 PM »
Don't the resistor strings need to add up to 100K?
No. The equation for bandpass center frequency is set by 1/(2PI * SQRT( (RFx.9+pot.value) * Cfb *  (RFx.10+pot.value) * Cshunt) ).
Keeping cap values ratio from Sontec circuit and setting the substituted pot.values to zero ohms (CW end of pot travel) giving the upper end of the bandpass filter.
Your FI wanted 5,500 Hz for the 2nd low mid band requires a RFx of 2,323 Ohms instead of Sontec 7,987 Hz with RFx 1,600 Ohms. Increasing the pot.value to 11,129 Ohms (giving series resistance 2,323+11,129 Ohms) brings the BP center down to your wanted 950 Hz for the substituted potsCCW end instead of Sontec 100,000 Ohms for 126 Hz.
Quote
It still doesn't make sense to me because the last switch position is 0Ω.
The CW end of the substituted stereo pot with rev.log taper is zero ohms, setting the upper end of the BP center frequency with the remaining 2,323 ohms series resistors.
A common mistake that people make when trying to design something completely foolproof is to underestimate the ingenuity of complete fools.  -Douglas Adams


Gold

Re: sontec 1 RU build thread
« Reply #761 on: January 31, 2019, 06:50:42 PM »
Your FI wanted 5,500 Hz for the 2nd low mid band requires a RFx of 2,323 Ohms instead of Sontec 7,987 Hz with RFx 1,600 Ohms. Increasing the pot.value to 11,129 Ohms (giving series resistance 2,323+11,129 Ohms) brings the BP center down to your wanted 950 Hz for the substituted potsCCW end instead of Sontec 100,000 Ohms for 126 Hz.The CW end of the substituted stereo pot with rev.log taper is zero ohms, setting the upper end of the BP center frequency with the remaining 2,323 ohms series resistors.

I think I'm starting to get it. To use the 11,129 series resistance calculated in the table I have to change both RF3.9 (1600Ω) and RF3.10 (1600Ω) to 2,323Ω.  To use the values from the series resistance table for the other three bands I have to change the RFx.9 and RFx.10 to the values in the top table. Lo = 4,260Ω LoMid = 14,200Ω  High = 2568Ω.

Harpo

Re: sontec 1 RU build thread
« Reply #762 on: January 31, 2019, 07:05:50 PM »
yepp (or change cap values)
A common mistake that people make when trying to design something completely foolproof is to underestimate the ingenuity of complete fools.  -Douglas Adams

Harpo

Re: sontec 1 RU build thread
« Reply #763 on: February 01, 2019, 03:17:54 AM »
Here is what I got empirically. Way different than the spreadsheet. I must not understand something.


Cap values for the Low band are 4.7uF / 3.3nF and 470nF / 330pF for the LowMid and HiMid band. Your measured 4.69uF parts value for the HiMid band is a decade high. Just sayin'...
« Last Edit: February 01, 2019, 04:00:55 AM by Harpo »
A common mistake that people make when trying to design something completely foolproof is to underestimate the ingenuity of complete fools.  -Douglas Adams

Gold

Re: sontec 1 RU build thread
« Reply #764 on: February 01, 2019, 04:13:25 PM »
Cap values for the Low band are 4.7uF / 3.3nF and 470nF / 330pF for the LowMid and HiMid band. Your measured 4.69uF parts value for the HiMid band is a decade high. Just sayin'...

It's a typo. The cap is 0.47uF. The first two measured values are in reverse order too.

I've ordered resistor values that I didn't have to make the switches. They should be here Monday.

I have a lot of 0.1% 10K resistors. Is there any reason not to change R1-R4 in the Input/Summing board to 10KΩ for improved CMRR?

Would I see better level accuracy by using 0.1% 20K Ω resistors for R18-21? The band level summing resistors. What about using 0.1% resistors for R8/R12 in the Output/Summing section?


Harpo

Re: sontec 1 RU build thread
« Reply #765 on: February 03, 2019, 05:25:29 PM »
I have a lot of 0.1% 10K resistors. Is there any reason not to change R1-R4 in the Input/Summing board to 10KΩ for improved CMRR?
IMHO no. Good move. I'd increase the 12pF for 663kHz LPF too optimistic compensation cap value of C57 (would rise to even more insane 1.3MHz LPF with your 10K R4) to maybe 68pF for 234kHz.

Quote
Would I see better level accuracy by using 0.1% 20K Ω resistors for R18-21? The band level summing resistors.
No, (the 100Ω resistors R9 and R15 already spoil the accuracy party by factor 5) and for your wanted +/-3dB boost/cut range, these 20Ks would increase by factor 4 anyway...
Quote
What about using 0.1% resistors for R8/R12 in the Output/Summing section?
Good move and decrease both to 10K in order to half the needed parts values of the summing resistors R18, R19, R20 and R22. Parts value of compensation cap C23 as described above.
As always YMMV.
A common mistake that people make when trying to design something completely foolproof is to underestimate the ingenuity of complete fools.  -Douglas Adams

Gold

Re: sontec 1 RU build thread
« Reply #766 on: February 03, 2019, 11:42:15 PM »
Thanks Harpo. With any luck I’ll have pictures in a couple of weeks.

Gold

Re: sontec 1 RU build thread
« Reply #767 on: March 07, 2019, 06:15:27 PM »
It's alive! It took close to 100 hours of bench time. This is a four channel EQ for vinyl mastering. It's for the A/B path console I'm building to master directly from tape to lacquer. All the PCB's are mounted on 3/8"x1/4" aluminum stock. The aluminum stock is mounted to the sides of the chassis. That way there is access from both the top and the bottom. I cut slots in the side of the case so the PCB's could slide back and forth for better access to the terminal blocks. The slots look really rough but when faced with the choice of spending fifteen minutes with an angle grinder and over an hour on the micro mill I went with the angle grinder.



View looking through the unit.



Starting to install the switches



Switches installed



I thought having all that braided shield packed in there might have a problem shorting to stuff. But I hoped against hope it would work. No dice. Have to put clear heat shrink over all the braided shield.



Time to put it back together again.



I had two channels that I put together with pots to work through all the choices. I worked out gain, frequency and Q ranges.

I built it like I think most others. I started out with OPA604/2604 OpAmps and Wima capacitors. When I started listening I wasn't happy. The EQ sounded hard and edgy. I figured I'd try swapping OpAmps and see if that helped. I tried some LME49710/20's. It helped a little but not nearly enough to turn my frown into a smile.

After thinking about it I figured the most likely culprits were the capacitors in the tuning circuit. I pulled the Wima's. I had some other brands of 0.47uF PP caps and some 330pF polystyrene caps. I redid the two mid bands and the improvement was dramatic. I now had a smile. I ordered a few different brands of PP caps to try for the larger value. I liked the polystyrene for the smaller value so I used what I had and ordered what I didn't.

Then I hand matched all the caps to within 1%. I of course had to order 3X what I needed to sort.



Everything tested with IC's. Time to put in the DOA's. I decided to run the filter boards +/-15VDC and the In/Out/ Summing boards +/-24VDC. The console also has separate PSU's for Preview and Modulation channels. The power distribution is from the grey terminal blocks in the rear.



Knobs on.



Listening in the studio.

« Last Edit: March 07, 2019, 06:48:41 PM by Gold »

scott2000

Re: sontec 1 RU build thread
« Reply #768 on: March 07, 2019, 07:16:38 PM »
Wow!
NIce!

john12ax7

Re: sontec 1 RU build thread
« Reply #769 on: March 08, 2019, 12:44:14 AM »
What PP cap did you end up with?


Gold

Re: sontec 1 RU build thread
« Reply #770 on: March 08, 2019, 01:21:53 AM »
What PP cap did you end up with?

The low end is a Cornell Dublier. The two mid bands are Illinois Capacitor and the high band is TDK.

I didn’t spend a lot of time on it. I think the polystyrene may be doing the heavy lifting.

The original swap was  two Wima MKS4’s for a Cornell Dublier and a polystyrene.

I wasn’t able to make the 3n3 in the low band a polystyrene until I figured out I could use three 5000pf in series parallel to get the value. The large 4.7uF made a difference but when I put the polystyrene in it helped a lot.

john12ax7

Re: sontec 1 RU build thread
« Reply #771 on: March 08, 2019, 01:41:47 AM »
Thanks for the info.  Wima are generally pretty good but I do remember MKS to be rather bad for audio.  The MKP series should be better,  with the FKP even better,  but limited in capacitance size.

Gold

Re: sontec 1 RU build thread
« Reply #772 on: March 08, 2019, 02:02:41 AM »
I don’t think the boards would take the footprint of an MKP. Since the leads are short you can’t shoehorn them in there. The smaller value was actually an MKP.

Edit: the smaller values were actually FKP to start with.
« Last Edit: March 08, 2019, 01:32:57 PM by Gold »

dagoose

Re: sontec 1 RU build thread
« Reply #773 on: March 12, 2019, 04:48:09 AM »
wow! That's a real DIY sontec there! Cool looking panel, suits it. Congrats.  8)
Da Goose Music Mastering http://www.dagoosemusic.nl

5v333

Re: sontec 1 RU build thread
« Reply #774 on: March 12, 2019, 06:19:06 AM »
applause!

Harpo

Re: sontec 1 RU build thread
« Reply #775 on: March 12, 2019, 07:16:18 AM »
Congrats, that's some serious DIY.
A common mistake that people make when trying to design something completely foolproof is to underestimate the ingenuity of complete fools.  -Douglas Adams

Gold

Re: sontec 1 RU build thread
« Reply #776 on: March 12, 2019, 10:52:31 AM »
Thanks guys. I don’t think I could have completed this build even a few years ago. 

I’m happy with the arrangement of controls. I oriented the controls vertically but in the horizontal order. It goes from left to right Lo Left, Lo Right, LoMid Left, LoMid Right, etc. It’s great for stereo or M/S but would be confusing for dual mono use.

I’m happy with the sound too.  I might call it YR-1 for Yaht Rock. It has that soft focus Super 8 kind of sound. That’s exactly what I was hoping to get out of it.

I’m working on the second unit for the console now. The only thing I’m going to change is some of the terminal blocks. I thought there would be better screwdriver access from the rear so I got angled blocks. It didn’t work very well so I’m using regular 90 degree blocks.

I put some shots of me making the faceplate in the Machine Shop.

Gold

Re: sontec 1 RU build thread
« Reply #777 on: March 12, 2019, 12:59:25 PM »
I had a happy accident when building this. The Electroswitch C4's have D shafts. The OKW knobs are collett mount. I've had both the switches and the knobs for years. If you just try to tighten the colletts on the shaft they slip. My first thought was to get some 1/4" half round stock to make the D into an O. The flat of the shaft doesn't take up 50% of the shaft diameter so it's not a drop in solution. The next thought was to get some thin wall brass tubing and glue it over the shaft. That didn't work either because no tube wall was thin enough to get the collett over.

When testing this EQ out I didn't want to take the knobs on and off a million times so I used a pliers with teeth to rotate the switches. It gave me a good grip but tore up the switch shafts. I figured I could smooth them out later if necessary. When I went to put on the knobs I discovered I had inadvertently made a lock washer. The colletts were snug over the roughed up shafts and don't move. Problem solved!


 

Related Topics

  Subject / Started by Replies Last post
128 Replies
74059 Views
Last post May 14, 2018, 09:48:30 PM
by vitopower
SSL THREAD + 1176 THREAD in 6 pages!!!!!

Started by gerardmanvuca « 1 2 » The Lab

21 Replies
6340 Views
Last post June 01, 2018, 01:03:57 PM
by SIXTYNINER
48 Replies
18045 Views
Last post February 20, 2019, 12:16:11 PM
by jensenmann
10 Replies
4099 Views
Last post September 28, 2010, 07:29:59 AM
by radiance