P2P Redd 47 - a few questions

[abbey road d enfer]

It just occurred to me that these modules must have had faders controlling their outputs in the desk. Why not just put a log pot on the output?

Although that would give you the possibility to decrease the overall gain, that would not increase the input headroom.

Or a combination of this an the 18dB input pad?

Some form of input pad seems to me to be the only reasonable solution to your concern.
You have the choice of a pre-xfmr pad, a posr-xfmr pad, that may me switchable or a pot, and the alternative of a custom-built input xfmr with a tapped or switched ratio (e.g. switching from series to parallels two identical sections would give 6dB attenuation without loss of quality).

 

[earthsled]

The like the idea of having a switchable input pad in addition to a continuous output level control.

What would be a good value for an output pot? I read that the original may have been a 200 ohm "T" attenuator. Those are difficult to find. Could a simple log pot be used instead? Also, would there be any need to use R23 hookup at the output?

For the input, an 18dB "H" pad can be made with 2K/200R impedances. I used an online calculator for the values attached...

 

[abbey road d enfer]

What would be a good value for an output pot? I read that the original may have been a 200 ohm "T" attenuator. Those are difficult to find.

Yes, everything was 200r at Abbey Road. But in fact the output of the Redd47 is quite tolerant to load variations, thanks to NFB.

Could a simple log pot be used instead?

I guess it would work. You would need a low value pot, in order to maintain a reasonably low impedance for the equipment it is connected to. Depending on cable length and capacitance and RFI environment, the value should be 1-5k. Beware that output CMRR will suffer from the non-symmetrical arrangement. Probably not an issue.

Also, would there be any need to use R23 hookup at the output?

No.

For the input, an 18dB "H" pad can be made with 2K/200R impedances. I used an online calculator for the values attached...

You don't want the 24r resistors. The lower the source impedance the better. U attenuator for mic inputs.

 

[Kamel]

I use a 1k log pot at the output, ala API, made several recording with no problem with the asymetrical arrangement

 

[earthsled]

Thanks Abbey, this is a great help. 

I guess it would work. You would need a low value pot, in order to maintain a reasonably low impedance for the equipment it is connected to.

If using a log pot, would it be connected across the secondary of the output transformer, creating a shunt resistance?

Depending on cable length and capacitance and RFI environment, the value should be 1-5k. Beware that output CMRR will suffer from the non-symmetrical arrangement.

I suppose the best way to avoid the CMRR issue would be to build a balanced attenuator on a rotary switch. Would a switchable "T" configuration be best for this?

U attenuator for mic inputs.

I calculate 698R for the series resistances and  200R for the shunt.

 

[abbey road d enfer]

Thanks Abbey, this is a great help. 

I guess it would work. You would need a low value pot, in order to maintain a reasonably low impedance for the equipment it is connected to.

If using a log pot, would it be connected across the secondary of the output transformer, creating a shunt resistance?

Yes; no problem, the 47 can drive 200 ohms load.

Depending on cable length and capacitance and RFI environment, the value should be 1-5k. Beware that output CMRR will suffer from the non-symmetrical arrangement.

I suppose the best way to avoid the CMRR issue would be to build a balanced attenuator on a rotary switch. Would a switchable "T" configuration be best for this?

Again I would suggest a U attenuator, because we don't need to maintain constant impedance.

U attenuator for mic inputs.

I calculate 698R for the series resistances and  200R for the shunt.

Seems correct. In the end it depends on the real impedance presented by the input xfmr primary and the accuracy you wish to attain. Considering the source Z presented by mics is a significant variable (from 100 to 600ohms), I wouldn't worry too much.

 

[earthsled]

I use a 1k log pot at the output, ala API, made several recording with no problem with the asymetrical arrangement

How do your have your 1k log pot connected? I guess I'm confused as to what "asymmetrical" means in this case.

Thanks!

 

[Kamel]

the two end on the output transformer (+ and -) the wiper is the output and referenced by the (-) of the transformer output

 

[earthsled]

I'm pretty sure Kamel is describing "A" -- just curious if "B" would work too.

Thanks!

 

[abbey road d enfer]

B would work somewhat, meaning it would certainly dim the output at some point, but that would be accompanied with a lot of distortion.
For a certain part of the rotation, the variable resistor would create a voltage divider with the output impedance of the preamp, and a t some point, the output stage will have to fight too much against the load and strat to break up.

 

[earthsled]

Thanks so much for all the help with developing a gain control! This was a great discussion, and I'm very pleased with the conclusion. I'm so glad we have Prodigy-Pro as a resource! You guys are great!  ;D   

Onward to transformers...

I've read about a few folks who have tried the Edcor 5:1 transformer as an inexpensive alternative for the output. I wonder if anyone has experience with the Hammond 143L? Mouser sells these for about $30(US) - which is much more affordable than the offerings from Cinemag and Sowter. The 143L seems to have the correct ratio of 10K to 200 ohms, but I'm not sure if the rating of "Output = 30mW" is sufficient for this circuit. Unfortunately, Hammond doesn't have many specs listed for this series, but I've attached the data sheet for what it's worth...

Also, I've stumbled upon what I think would make a good choice for a dual-channel power transformer. Weber VST makes replacement transformers for guitar amps, and their W025130 or W025130EU provides necessary taps for 540V@150mA and 6.3V@5mA. In addition, this transformer has an extra 45V tap that would be useful for making a 48VDC phantom power supply! The W025130 / W025130EU is priced at $50 / $54 respectively which is a bit more than the Allied 227-0081, but with the extra tap for phantom, I think it's definitely worth the extra $4 to $8.

Here's the link to Weber's transformers... https://taweber.powweb.com/store/magnetic.htm

I'm planning to mount the PT outside the chassis through the back panel - similar to the Manley Vari-Mu compressor...
http://www.sweetwater.com/images/closeup/xl/1600-VariMu_rear.jpg

Cheers!

 

[abbey road d enfer]

I've read about a few folks who have tried the Edcor 5:1 transformer as an inexpensive alternative for the output. I wonder if anyone has experience with the Hammond 143L? Mouser sells these for about $30(US) - which is much more affordable than the offerings from Cinemag and Sowter. The 143L seems to have the correct ratio of 10K to 200 ohms, but I'm not sure if the rating of "Output = 30mW" is sufficient for this circuit. Unfortunately, Hammond doesn't have many specs listed for this series, but I've attached the data sheet for what it's worth...

It just doesn't work. 30mW into 200r is less than 2.5 Vrms or +10dBu. Not much headroom. But wait, there's more; this power is for aN LF corner frequency of 150 Hz. When you derate at 12dB/octave, you get about -25dBu at 20Hz! I think it's not even worth trying. Take the Sowter, you won't be disappointed; these guys know what they're doing.

Also, I've stumbled upon what I think would make a good choice for a dual-channel power transformer. Weber VST makes replacement transformers for guitar amps, and their W025130 or W025130EU provides necessary taps for 540V@150mA and 6.3V@5mA. In addition, this transformer has an extra 45V tap that would be useful for making a 48VDC phantom power supply! The W025130 / W025130EU is priced at $50 / $54 respectively which is a bit more than the Allied 227-0081, but with the extra tap for phantom, I think it's definitely worth the extra $4 to $8.

You'll have plenty of power with this.

 

[earthsled]

It just doesn't work. 30mW into 200r is less than 2.5 Vrms or +10dBu. Not much headroom. But wait, there's more; this power is for aN LF corner frequency of 150 Hz. When you derate at 12dB/octave, you get about -25dBu at 20Hz! I think it's not even worth trying.

Because of my budget, I'd still like to pursue a less expensive alternative for the output.
Edcor makes two models that may work:

Edcor XSM15K/600
5:1 at 15K/600
http://www.edcorusa.com/products/164-xsm15k-600.aspx

Edcor XSM10K/150
8.2:1 at 10K/150
http://www.edcorusa.com/products/159-xsm10k-150.aspx

Which one would be the better choice? Would changes to the circuit need to be made to use these different ratios?

If anyone knows of other affordable options please post!

 

[Kamel]

Sorry to reply so late, the "A" version, 1k log, as done in some API stuff, not theorically perfect, but with the actual converter with 10k input, it's ok

 

[earthsled]

Sorry to reply so late, the "A" version, 1k log, as done in some API stuff, not theorically perfect, but with the actual converter with 10k input, it's ok

This seems like a good solution to me. Have you experienced any issues when using this output patched into something like an LA-2A or other gear with less than 10K impedance?

Thanks!

 

[Kamel]

never experiment with 600 load, because i never nedded it, but theoricallly, as the 47 can drive 200 it's not a problem except the log curve of the pot will change, and will not stay log, to resolve this use a switch to return as fixed output and use the input level of the gear behind (LA2 or everything 600 load)

......

i made some test with the 5:1 transformer , with a 10k load it is ok, but when the load decrease, as expected, distortion and saturation increase, with a 8:1, what i have on hand, it's perfect, can drive a low Z without problem

 

[earthsled]

Kamel -- Thanks for the switch idea -- that is a great way to handle the issue.

I have a question about the caps in the power supply...

Does it make any difference between having 3 x 8uF caps in parallel vs. 1 x 24uF? Is there any advantage to using the separate caps?

Thanks!

 

[abbey road d enfer]

You can use a single 22uF in place of 3x 8uF. Doesn't make a ghost of a difference.

 

[earthsled]

With regards to grounding...

The P2P layout and the schematic appear to have two different ground sections (G1 and G2). Should these be tied together? Should either one be connected to the chassis?

I've often seen equipment with connections for "chassis ground" and "analog ground" on binding posts. Does the "analog ground" typically include the DC side of the power supplies?

Thanks!

 

[abbey road d enfer]

With regards to grounding...

The P2P layout and the schematic appear to have two different ground sections (G1 and G2). Should these be tied together?

Indeed they must.

Should either one be connected to the chassis?

Yes, although there are some options there. Some choose to connecte the audio ground to the chssis ground via an RC network constituted of a resistor of 10r to 1k in parallels with a cap of 0.01 to 1U. This is not necessary on a unit that has balanced in/outs, but you may want to make provision for that option by connecting the audio ground rail and the chassis ground to two adjacent lugs of a terminal strip; then you may experience with differnt RC values or a plain and simple wire. This should not be necessary in your case, and may not make much difference, but it is easier to have that  option rather than having to modify the ground rail.

I've often seen equipment with connections for "chassis ground" and "analog ground" on binding posts.

That's another option for decoupling audio gnd and chassis gnd.

Does the "analog ground" typically include the DC side of the power supplies?

Yes.