Another vari-mu: Chiswik Reach Compressor

[Viitalahde]

Yep, nice drawing. Thanks! :thumb:

 

[synthi]

Yes! Thanks Olaf.

I just made some more hi-res closeup photos and spoted a couple of values, the Olaf work is great.

What I`m not sure about is the input transformer, I was told than in early version they used a small trafo (and seems to be the case with my VK1) that distorted with hi level signals.
If I send a 1Khz to at 0dBFS (+22dBm), with the input att engaged (-10dB) I need to set the gain pot (an attenuattor really, right??) past "6" I got nasty distortion, So with a quick calculations the thing is distorting with a +8dBm signal or so.
I think they used the 3575 in later units, so I think would be a good idea changing the original trafos with them...

Also, the input attenuation would be improved, any clues?

Also Looking to see your phoenix photos!

Thanks!

Synthi.

 

[synthi]

I just emailed sowter support about the input trafos for the chiswick. I`m pretty sure the input trafo in mine is not a 3575, maybe same ratio but inferior specs ordered as a custom job for early works, and then sowter comes with the improved 3575 and they are using it now.

Would be interesting to know what trafos are used in the phoenix too.

I`ll let you know when I get a reply from Sowter...

Best Regards,

Synthi.

 

[synthpunk]

the vk1 stopped using sowters years ago... they were replaced with massive, much higher quality ones, i think possibly sourced from Stevens and Billington.anything with sowters in it is either a prototype or so early as makes no difference- we were learning from every one we built! the circuit linked to above must be an early one, cos it is definitely different now... and its no longer true point to point- theyre now built on 'voyager board' - the dual row, 36 tag phenolic board from RS that the voyager space probe was built on.... the horrible birds nest point to point style was used on the first dozen or so.....

i know this cos i used to build vk1's when they were made in chiswick, after vic left, and before they started getting them built elsewhere.

i agree that it is a very pretty drawing! its nice to finally see it out there, ive been sitting on the schem for years, but i promised nigel (the guy who owns CR) that id never give it to anyone else...

a fine piece of reverse engineering! well done, olafmatt..

cheers,

rrrrich

 

[synthi]

Hi Synthpunk,

CW used the sowter for a bit more than a dozen for sure... Mine is #14 :wink:
Are they now using different output trafo or just changed the input? I think I read somewhere just the input trafo was upgraded...
Also, can you remember the specs and ratios for the trafos used in later revisions, Also, do you remember what was changed in later revisions? You know I`d like to upgrade mine...

Thanks,

Synthi.

 

[synthi]

Hi People.

I did some research about the Vk1:

I emailed sowter about the trafos and they are really helpful! Brian Sowter spoke with Vic about the VK1, great. :grin:

Brian told they was dicussing about the saturation and the new trafos would be an improvement , but not a solution... I know understand why:

I was checking the vk1 scheno vs the Altec and they are very similar. The VK1 is having the same behaviour than the Altec 436C, the stock -10dB pad is no enought for hot signals, and the output is really hot, so I`m planning to add a 600omh stepped attenuator at the input and output. One of the things used in the pheonix compressor and later Chiswik models is having switchable 3 position pads for inputs and outputs, I think for correcting this.

Also looking how the input gain is implemented in the 436C and the Chiswik, the altec have the dual 50K pot before the trafo, but the vk1 have the dual 47K pot AFTER the trafo. It seems to be very crude in both designs, I`d like to have more precise gain control. Is there a more elegant solution than using dual pots here? NYD? CJ?

Thanks!

Synthi

 

[Joechris]

The Altec pot is also after the trafo. Use a resistor string pad instead.
Look at Davids Altec mod thread somewhere.
This is a great calculator:
http://homepages.tcp.co.uk/~nroberts/atten.html
jo

 

[Joechris]

So with a 2pole 12 way rotary, it could look like these.

Step 1, Attenuation = Infinity dB, Rx = 50000 ohms, Ry = 0 ohms, Resistor = 0 ohms.
Step 2, Attenuation = 35 dB, Rx = 49111 ohms, Ry = 889 ohms, Resistor = 889 ohms.
Step 3, Attenuation = 28 dB, Rx = 48009 ohms, Ry = 1991 ohms, Resistor = 1102 ohms.
Step 4, Attenuation = 24 dB, Rx = 46845 ohms, Ry = 3155 ohms, Resistor = 1164 ohms.
Step 5, Attenuation = 21 dB, Rx = 45544 ohms, Ry = 4456 ohms, Resistor = 1301 ohms.
Step 6, Attenuation = 18 dB, Rx = 43705 ohms, Ry = 6295 ohms, Resistor = 1839 ohms.
Step 7, Attenuation = 15 dB, Rx = 41109 ohms, Ry = 8891 ohms, Resistor = 2596 ohms.
Step 8, Attenuation = 12 dB, Rx = 37441 ohms, Ry = 12559 ohms, Resistor = 3668 ohms.
Step 9, Attenuation = 9 dB, Rx = 32259 ohms, Ry = 17741 ohms, Resistor = 5182 ohms.
Step 10, Attenuation = 6 dB, Rx = 24941 ohms, Ry = 25059 ohms, Resistor = 7318 ohms.
Step 11, Attenuation = 3 dB, Rx = 14603 ohms, Ry = 35397 ohms, Resistor = 10338 ohms.
Step 12, Attenuation = 0 dB, Rx = 0 ohms, Ry = 50000 ohms, Resistor = 14603 ohms.

 

[synthi]

sorry I reversed the facts: the altec pot is after the trafo and the vk1 is before the trafo :wink:

So I`m thinking now: input signal-->stepped attenuator-->variable T attenuator would be the solution?

ITOH, Thanks for the link for the calculator!

I see you post values for a 50K attenuator , its better than using 600ohm for both inputs and outputs?

Also I think I`ll need Balanced stepped attenuators and balanced variable T attenuator, right?

Thanks a lot friends!

Synthi.

 

[synthi]

I think the output attenuator was in off position, but I`ll do some quick tests again....

More on this later today...

Synthi.

 

[synthi]

Ok, tested:

Now I know the facts about the altec 436 operation, I`m feeding the unit with a -22dBFS (no pad) and -12dBFS (padded) and is working as it must.
I must pad the output at certain settings because it will over my AD sometimes, lot of gain inside really...

With an input at -22dBFS (0dBm), no pads, release to slowest, attack to fastest and threshold to "1" (minimun) I get unity gain with input level control at "3" (marked 1 to 11) and I get a digital over when level at " 9.8" but output is distorting when the level pot past "6" (about -6dBFS at the AD, compression meter reads -3.5dB compression). Now the funny part: If I switch the OUTPUT pad "on", then I can go thru the entire range up to "11" without the distortion... (level "11": -11dBFS at AD, comprression 10dB)

I`m using Apogee DA16--->VK1--->Mytek AD, all is calibrated and I tested: conecting the DA16--->Mytek AD, the ad converter is not distorting up to -0.04dBFS out from the DA16, so the AD is not overloading for sure. I`m using the RME digicheck for checking the inputs at the AD.

Its strange (at least for me) that the distortion goes by when I switched the output pad now that is clear that is not the AD who is distorting, so I think It could have something to do with output impedance being "better" when switching the output pad?
Also I noticed some phase isues, I think that would be becuase the input pots having variable impedance and the way attenuators are implemented... (?)

Also I have some thoughts about the best values for attenuation:
input attenuator: -22dB, for having a 0dBFS signal into 0dBm optimal operating level, or better having several values: -30 (for using the compressor as gain amplifier for coloring without compression) -22dB and -8dB for prosumer gear.

output attenuator: here the most steps the better for precise control, 6 steps is ok for interfacing with the AD or Professional gear. Seeing my tests, having 6dB attenuation would be enought for gear acepting +22dB signals, so 1 to 6 dB attenuation would be great, a better option would be 12 steps between 1 to 8dBs since some called pro gear can handle +20 max.

What do you think about this?

Synthi

 

[synthpunk]

both input and output TX's were changed- the new ones are about the size of a grapefruit, so a good deal bigger than the lychee sized original....

the original ones distorted on the low end, thats why they were changed.

and yer right, the first 20-30 or so had little tx's, not the first dozen.... if i remember right, i started building em from about number 18ish, till about number 30 ish, as a summer job between the 2nd and 3rd year of uni in the summer of 97..all the ones i built had little sowter tx's... some time after that they started being made out of house, somewhere in holloway rd, london.. i think they must have changed the trafos around the same time... chiswick reach studios themselves closed down in '99/2000, so its only the compressors left nowadays.. shame, it was a great place to record, and work, too!....

the output attenuator was also changed on the later ones, with more attenuation settings..


rrrrrrrrrrich

ps

 

[NewYorkDave]

It gets awfully confusing when you mix dBFS and dBM together like that, especially since everyone seems to have his own idea of how many dBM (or dBU) equals 0dBFS. The practice I follow is maximum level=20dB above standard operating level, therefore 0dBFS=+24dBM where S.O.L. equals +4dBM. But I know that there are others who only consider 18dB of headroom to be necessary, instead of 20.

I recommend that for the purposes of this discussion, since we're talking about an analog device, we stick to dBM (or dBU if you prefer).

 

[Larrchild]

To add to the confusion, the headroom @50 hz is different than the 20 hz spec. Sowter's site says to subtract some db's @ 20 hz.

 

[synthi]

The practice I follow is maximum level=20dB above standard operating level, therefore 0dBFS=+24dBM where S.O.L. equals +4dBM. But I know that there are others who only consider 18dB of headroom to be necessary, instead of 20.

I know lot of people from USA use the smpte standard of 0dBFS=+24dBM , we EU people mostly use 0dBFS=+22dBM

Sorry if if its confusing....

Synthi

 

[Lukpio]

BTW if I remember correctly, ADATs used +18dBu as 0dBFS and old Mitsubishi digital multitracks used +19dBu=0dBFS.

 

[bernbrue]

Hi Synhi,
I followed this thread for a quite long time now and I think the VK1 could be a very good DIY project. Here are a few questions:

Did you manage to repair and setup your VK1?
Is the schematic that was published here correct, does it work?
Have you got further information of this beast?

Maybe we can start a nice DIY project out of it. Frank (nrgrecording) has already made a pcb layout from your schematics. I'm just making a BOM for this project.

Regards
Bernd

 

[bernbrue]

Hi Olaf,
I would like to invite you to our German forum. You may contribute a few things to our thread. Look here please:
http://techtalk.te.funpic.de/index.php?topic=166

See you
Bernd

 

[gary o]

I currently fiddling with this circuits 6AL5 side chain adding it to a BA6A, its working nice giving me more control over the limiting.

Question what is the thumps switch meant to do ??????????????
Is that switch on one of the pots ????? and why ????????


Thanks people........................Gary O.

 

[synthi]

hi Bernbrue,

Yes, I have my vk1 woring, and sound awesome!

I was driving the unit too hard, lower levels ans all is fine. This is because the trafos used, later they changed the trafos for hotter levels, but I`m also recording lower levels now so is not a problem anymore...

Also as olaf pointed, they changed the tubes to ecc85.

I kweep mine unmodded but would be great DIYing another one for playing with.

The thump switch engage a very slow attack, usefull for percussive sounds.

I`m sorry I can read german so would be great if you can move the Vk compressor here again!

Hare you planning to fab the PCBs? NGR where are u!? :thumb:

Synthi