REDD EQ, Helios 69 and Dick Swettenham

[tommypiper]

You could add a switch to connect the input of the second stage of the classic either to the EQ output or to a line input transformer. The EQ input would then be unbalanced high impedance (just what the EQ likes) and the line in would be 10K balanced.

Excellent, that's what I was thinking.  Will the Classic 10k attenuator and HPF work with both the line input (with input transformer) and the unbalanced EQ input?  Does that change the load on the EQ from 1M to 10k?

And no problem connecting the EQ to Classic with unbalanced cable...?  Use screened cable, screen connected to one end...?

 

[salomonander]

Nice build!  Can you share how you are wiring the transformers?  Which carnhills are you using on the outputs?

i wont comment since it doesnt work at this point. i use a 15k:600 carnhill. in theory it makes the output look like 200ohms in order to feed it into a micpre. but it picks up crazy amounts of hum at this point. i have not found a solution so far. i just ordered a mu metal can for the output xformer. i hope it helps. otherwiese im out of luck and will most likely have to fit some internal makeup.

 

[ruffrecords]

Excellent, that's what I was thinking.  Will the Classic 10k attenuator and HPF work with both the line input (with input transformer) and the unbalanced EQ input?  Does that change the load on the EQ from 1M to 10k?

No, it will work with the input transformer because as you have guessed it changes the EQ load to 10K which is not what we want.

And no problem connecting the EQ to Classic with unbalanced cable...?  Use screened cable, screen connected to one end...?

Screened cable should be fine provided it is no too long (say 1 metre or less). Connect screen at both ends to 0V.

Cheers

ian

 

[tommypiper]

No, it will work with the input transformer because as you have guessed it changes the EQ load to 10K which is not what we want.

I meant, will it work without an input transformer in the Classic, but with the 10k attenuator... I believe the answer is the 10k attenuator will create a 10k load.  So, how do I control gain there without loading the EQ?  Can we use a 100k attenuator? 

Since the circuit load is 1M, would it make sense to make a 250k attenuator?  And incidentally, will the Classic HPF (in your instructions which needs a 10k load) work ok with the higher Z?

thanks!

BTW:  AML is not listing one of the Helios inductors:  VTB9042.  Is there another source?

 

[ruffrecords]

I meant, will it work without an input transformer in the Classic, but with the 10k attenuator... I believe the answer is the 10k attenuator will create a 10k load.  So, how do I control gain there without loading the EQ?  Can we use a 100k attenuator? 

100K should be OK. It is high enough not to load the EQ but low enough not to br particularly susceptible to interference.

Since the circuit load is 1M, would it make sense to make a 250k attenuator?  And incidentally, will the Classic HPF (in your instructions which needs a 10k load) work ok with the higher Z?

I think there would be no additional benefit in going to 250K. If you use 100K then you will need to use different capacitors for the HPF. Since 100K is 10 times 10K, the capacitors need to be reduced 10 times for the frequencies to remain the same.

BTW:  AML is not listing one of the Helios inductors:  VTB9042.  Is there another source?

That's very odd. I do not know of another source other than Carnhill themselves. I will ask Colin.

Cheers

Ian

 

[salomonander]

BTW:  AML is not listing one of the Helios inductors:  VTB9042.  Is there another source?

i might have some for sale. pm me if needed

 

[ruffrecords]

BTW:  AML is not listing one of the Helios inductors:  VTB9042.  Is there another source?

I asked Colin at AML about this. Apparently someone bought all his stock of these. He has more on order but they won't be available for four weeks.

Cheers

Ian

 

[salomonander]

i have 7 for sale in case someone needs them

 

[EWalter]

Hey All,

Reviving this thread because I just built a REDD EQ and have a few questions. I'm planning to build it into a channel strip with an ADM-1668 card (Schematic attached). I'm curious if anyone has any recommendations as to where I should wire it in this signal path. I'm still trying to wrap my head around how and why impedances changes in circuits and I know it's important for this EQ to have a high impedance source.

The input xfo is a UTC 0-8 and is wired secondary to primary, so the input is 500 ohms and the output is 15k.

My stupid question is as follows: Does this mean that the impedance is 15k until it hits the output xfo? Or does it change with every component in the circuit? If so, should I add something to get the optimal impedance for this EQ?

I'm of course very interested in a simple answer of where I should insert the EQ in the signal path, but I'd also like to learn a little so I don't have to bug y'all in the future. Thanks in advance.

Cheers,
Erik

 

[ruffrecords]

Hey All,

Reviving this thread because I just built a REDD EQ and have a few questions. I'm planning to build it into a channel strip with an ADM-1668 card (Schematic attached). I'm curious if anyone has any recommendations as to where I should wire it in this signal path. I'm still trying to wrap my head around how and why impedances changes in circuits and I know it's important for this EQ to have a high impedance source.

It is important for my REDDEQ design to have a high impedance load (100K or higher if possible). It also works best with a source impedance of around 2500 ohms but it will work with any source impedance lower than this.

The input xfo is a UTC 0-8 and is wired secondary to primary, so the input is 500 ohms and the output is 15k.

I am not sure why you are using this at the input. The O-8 is an output transformer.

My stupid question is as follows: Does this mean that the impedance is 15k until it hits the output xfo? Or does it change with every component in the circuit? If so, should I add something to get the optimal impedance for this EQ?

I'm of course very interested in a simple answer of where I should insert the EQ in the signal path, but I'd also like to learn a little so I don't have to bug y'all in the future. Thanks in advance.

Cheers,
Erik

My REDDEQ looks like a load of between 5K and 10K depending on the setting of the controls and the frequency. The input transformer will reflect this impedance to the primary and hence the input. My REDDEQ is designed to be fed via a 10K:10K transformer so its input impedance is reflected more or less intact to the primary. Any professional audio source will then drive this input. So in this instance the input transformer really should be changed to a 10K:10K type. The EQ itself should be placed between the output of the transformer and the input of the op amp.

Cheers

Ian

 

[EWalter]

Thanks so much for the info!

It is important for my REDDEQ design to have a high impedance load (100K or higher if possible). It also works best with a source impedance of around 2500 ohms but it will work with any source impedance lower than this.

So it will work with anything, but works best 25k. What does "best" mean in this case? The curves will be closer to the original?

I am not sure why you are using this at the input. The O-8 is an output transformer.

Thats the way the card came. I attached the promotional material for the product.

My REDDEQ looks like a load of between 5K and 10K depending on the setting of the controls and the frequency. The input transformer will reflect this impedance to the primary and hence the input. My REDDEQ is designed to be fed via a 10K:10K transformer so its input impedance is reflected more or less intact to the primary. Any professional audio source will then drive this input. So in this instance the input transformer really should be changed to a 10K:10K type. The EQ itself should be placed between the output of the transformer and the input of the op amp.

I inserted it between the xfo pin 7 and R9 and it seems to be working well except the brilliance setting doesn't appear to be doing anything. I'm going to double check all my work, but I'm curious if this could have something to do with not having the ideal 25k source.

Regarding the 10K:10K: xfo, would you recommend that I add one to this circuit on the input of the REDDEQ?


Cheers,
Erik

 

[ruffrecords]

Thanks so much for the info!
So it will work with anything, but works best 25k. What does "best" mean in this case? The curves will be closer to the original?

It will work with any source impedance from zero to 2500 ohms. At 2500 ohms the curves will be closest to the original. The reason for this is the EQ is designed to be used in a mixer directly after a 10K fader. The source impedance of such a fader typically varies from zero to 2500 ohms. In a stand alone EQ, where the source impedance is likely to be close to zero, there is provision on the PCB to add a series resistor to raise this to 2500 ohms. This is covered in the documentation for the EQ.

Thats the way the card came. I attached the promotional material for the product.

Unfortunately the promotional info makes no mention of the transformers. It seems odd that the input transformer  position should be fitted with an output type, particularly as that type is also intended for use in tube output stages.

I inserted it between the xfo pin 7 and R9 and it seems to be working well except the brilliance setting doesn't appear to be doing anything. I'm going to double check all my work, but I'm curious if this could have something to do with not having the ideal 25k source.

That looks OK to me. Any wrong impedances are likely to affect all bands in much the same way. However,

Regarding the 10K:10K: xfo, would you recommend that I add one to this circuit on the input of the REDDEQ?

Cheers,
Erik

I do. I think you should replace the existing input transformer with a 10K:10K type.

Cheers

Ian

 

[Winston OBoogie]

I was recently asked if I could design an EQ that worked like the 'pop' and 'classic' EQ plug ins that used to be used in the REDD47 consoles and predecessors. Using the curves published in 'Recording The Beatles' I came up with a circuit based on a stripped down and modified Helios 69 EQ with a switch to selec 'pop' or 'classic'. I was then asked if I could not make it so both the the 'pop' and 'classic EQ curves were available at the same time. In doing this, a very strange thought occurred to me. The Helios treble EQ is virtually identical to the EMI 'classic' EQ curves (the frequency, step size and gain range are identical) as is the bass cut except for the frequency it works at. The bass boost is very nearly the same but tweaked from a shelf to a bell curve. The clincher is that the 'pop' 4.7KHz peaking EQ is a stepped version of one of the Helios 69 mid boost frequencies.

I then realised that Dick Swettenham, who designed the Helios 69 EQ, had previously worked at Abbey Road studios in the service and design departments so he must surely have had a deep understanding of the innards of the REDD EQ.

I am sure you can now see where I am going with this. Is the Helios 69 EQ simply a modified and expanded version of the REDD EQ?

I simulated a cut down version of the Helios 69 EQ (pic attached)  and it is surprisingly easy to get curves very close to those of the REDD EQ.

What do you think/know??

Cheers

Ian

Hey Ian,
I only just saw this.  Well done.  I did similar a few years back after plugging the original REDD eq schematics into spice and tweaking things to match using a similar topology to yours.
The original is a very well designed circuit but it isn't really practicable or necessary to use the original topology of two 400 ohm constant impedance stacked circuits for a  200 ohm with 10dB total loss -  rather than the usual (at the time) 10dB loss per band (+20dB total). 
For the 'Pop' presence and the brilliance controls, using a Pultec/Helios type circuit, I found it necessary to adjust the Q slightly for each step to match the original topology but I doubt it made much difference to what was heard/ perceived.

Regarding Dick Swettenham and EMI, I'm not sure there is any connection with the circuits.  The approach is very different, even if similar results can be obtained by manipulating things.  The REDD eqs were based on an earlier EMI desk eq that had  2OdB insertion loss which, in turn, was pretty much based on original research by Alan Blumlein and others. 
The Helios EQ was simply a lesser expensive to build passive LCR EQ when constant impedance wasn't needed.  It was quite commonly used before everyone adopted Peter Baxendall's feedback type EQ.  The Sound Techniques original A Range desks at Trident et al.  used a similar EQ to the Helios before switching to Baxendall type for the System 12 etc. 

That's all I can think of at the moment.
Cheers.

 

[EWalter]

That looks OK to me. Any wrong impedances are likely to affect all bands in much the same way.

I'm still having issues with the Brilliance band. High and Low bands are working great, but the brilliance has an extremely wide Q on all frequencies. They're so wide they pretty much just turn up/down the entire signal. I've double-checked all cap and resister values, beeped out each signal path to ensure I'm getting signal everywhere it's supposed to be and everything looks normal. I haven't tried the 10k:10k transformer yet, but everything else seems to be working great so I'm not sure that's the issue. I'm wondering if the inductor could potentially be bad...? Is there is simple way to test this and is it possible to do without pulling it out of the circuit?

I'm using the wire link instead of a 2k4 resistor for R35 because my transformer coming into it is high impedance. I believe that I'm thinking about this correctly. The manual mentions If a low impedance source is used and R35 is a wire link, then the Q of the brilliance curves will be sharper and the maximum boost and cut will be increased.. I have very wide Q's, so I'm not exactly sure how to proceed here.

Thanks again for all the help!

 

[ruffrecords]

I'm still having issues with the Brilliance band. High and Low bands are working great, but the brilliance has an extremely wide Q on all frequencies. They're so wide they pretty much just turn up/down the entire signal. I've double-checked all cap and resister values, beeped out each signal path to ensure I'm getting signal everywhere it's supposed to be and everything looks normal. I haven't tried the 10k:10k transformer yet, but everything else seems to be working great so I'm not sure that's the issue. I'm wondering if the inductor could potentially be bad...? Is there is simple way to test this and is it possible to do without pulling it out of the circuit?

I'm using the wire link instead of a 2k4 resistor for R35 because my transformer coming into it is high impedance. I believe that I'm thinking about this correctly. The manual mentions If a low impedance source is used and R35 is a wire link, then the Q of the brilliance curves will be sharper and the maximum boost and cut will be increased.. I have very wide Q's, so I'm not exactly sure how to proceed here.

Thanks again for all the help!

The Brilliance band does have a wide Q. The Q is approximately one. So on the 2K7 band for instance the -3dB points will be about 1K3 and 4K respectively. Attached is the plots for the 2K7 band as obtained from the official EMI plug in which is what I attempted to emulate in hardware.

Cheers

Ian

 

[Winston OBoogie]

The Brilliance band does have a wide Q. The Q is approximately one. So on the 2K7 band for instance the -3dB points will be about 1K3 and 4K respectively. Attached is the plots for the 2K7 band as obtained from the official EMI plug in which is what I attempted to emulate in hardware.

Cheers
Ian

And those curves look pretty much "spot on"

DCR of EWalter's  inductor could be widening  Q maybe?

Side note: in the early iterations of the EMI 'pop" EQ, the dcr of the low end inductor had been neglected in the design,  There was a production modification note added later which advised changing the lower end values of the bridge-T resistive networks to correct for it.

 

[EWalter]

Ok - I'm making some serious headway. I don't have any great analysis tools, so I've been struggling a bit up until now. I've been using pink noise and a spectrum analyzer which has been difficult to tell exactly what is going on, but today I used Plugin Doctor's linear analysis and got some very informative readings!

The Brilliance band does have a wide Q. The Q is approximately one. So on the 2K7 band for instance the -3dB points will be about 1K3 and 4K respectively. Attached is the plots for the 2K7 band as obtained from the official EMI plug in which is what I attempted to emulate in hardware.

So at this point everything is looking pretty spot on and inline with all of the plots included with the instructions, however the 2k7 band is still being problematic. It has an extremely wide boost and a very narrow cut  :'(

Everything else is looking fantastic - this is my last hurdle.

Below is a link to the screenshots of the measurements I got for the 2k7 cut/boost. I couldn't overlay them into one image, so I had to include a separate shot of each switch position:
https://imgur.com/a/mIAZVnk

Thanks again for all the help!

DCR of EWalter's  inductor could be widening  Q maybe?

This is a little over my head, but lemme know if the screenshots might set off any light bulbs. If there is something I should test on the inductor, I'm game!

Cheers,
Erik

 

[Winston OBoogie]

This is a little over my head, but lemme know if the screenshots might set off any light bulbs. If there is something I should test on the inductor, I'm game!

Cheers,
Erik

The birth father of this project - Ian - would be a better person than I am to answer specifics regarding q etc & I'm sure he'll be along soon. 
I don't have his latest/last revision schematic on hand but, looking at the boost plots you made @2K7, is it possible there's a short on the last winding of your inductor? 
I'm not exactly sure why the cut would be off in Q, it's using a different inductor than boost and is located as a shunt reactance.  But it could be widened by inserting some resistance in series with that particular tap of the inductor.  Maybe wait for Ian before changing anything there though. 
Check the boost inductor tap for problems there in the meantime.
Persevere, you're almost there! 
Whereabouts can I find Ian's up to date schematic?

 

[ruffrecords]

Several points:

1. You should use white noise rather than pink to look at EQ spectra.
2. The brilliance boost and cut Q should be virtually identical. The circuit uses separate LC circuit for boost and cut in order to do this
3. The Q of the circuit depends on the characteristic impedance which is made up of various circuit resistances and the dcr of the inductor.  The recommended inductors are chosen such that the contribution of the inductor dcr to Q is largely swamped by that of the other circuit resistances. The fact you are getting a sharper Q for cut compared to boost indicates one of them is wrong somewhere. I don't know if you are using one of my PCBs but it you are you should makes sure the resistor ladder values are correct and the frequency band select switch is wired correctly.

Cheers

Ian

 

[Winston OBoogie]

The recommended inductors are chosen such that the contribution of the inductor dcr to Q is largely swamped by that of the other circuit resistances. The fact you are getting a sharper Q for cut compared to boost indicates one of them is wrong somewhere.

I always assumed your design was 100% spot on Ian and you'd accounted for everything.  Hope it didn't sound like I thought otherwise.
I figured it was an issue with a wrong resistor value, and short or wiring issue on the boost inductor etc. 

Good luck EWalter