Hello everyone,
I'd like to remove the input stage from a crystal semiconductors cs5361 analog to digital converter evaluation board and use transformers instead.
the current/recommended stage is here:
http://www2.freepichosting.com/Images/34475/7.jpg
they used ne5532 opamps on the eval board, and a pathetic clock circuit, just 2 caps and a crystal. eeeeeeew.
I've got some jensen transformers i'd like to use instead for the input stage.
Has anyone got any experience with transformers on the front of ADC's, and or recommendations for this particular one.
Thanks,
Mark
I'd like to remove the input stage from a crystal semiconductors cs5361 analog to digital converter evaluation board and use transformers instead.
the current/recommended stage is here:
http://www2.freepichosting.com/Images/34475/7.jpg
they used ne5532 opamps on the eval board, and a pathetic clock circuit, just 2 caps and a crystal. eeeeeeew.
I've got some jensen transformers i'd like to use instead for the input stage.
Has anyone got any experience with transformers on the front of ADC's, and or recommendations for this particular one.
Thanks,
Mark
Are you looking for something.....
...like this?
...like this?
An externally hosted image should be here but it was not working when we last tested it.
http://f3.grp.yahoofs.com/v1/QK-zPjGBqXbkkSQMiMPv_2o__f0ac0KH6yOb8cRET7KRpK3fjPxkKhQqIiMkgcZx6to6DbEGpq94CRJ3/TL%20Circuits/Ultracurve/The%20Ironimous%20Solution%20for%20Behringer%20UC%208024.jpg
Koinichiwa,
It's fine from here, so no idea why it ain't working for you. Try instead:
http://groups.yahoo.com/group/Thunderstone_technical/files/TL%20Circuits/Ultracurve/
The file is called:
The Ironimous Solution for Behringer UC 8024.jpg
Sayonara
apassgear said:
It's fine from here, so no idea why it ain't working for you. Try instead:
http://groups.yahoo.com/group/Thunderstone_technical/files/TL%20Circuits/Ultracurve/
The file is called:
The Ironimous Solution for Behringer UC 8024.jpg
Sayonara
Koinichiwa,
The +15dbu are of course a type. You are right, it's +19dbu.
It works like this:
Vin for ADC full scale = 4V P-P for the CS5394 which is used in the Behringer EQ.
The transformer is 5:1 stepdown thus for full scale you need 4V * 5 = 20V P-P. Now 20V P-P = 7.07V RMS. And 7.07V RMS is appx. +19dbu.
Sayonara
Steve Eddy said:
The +15dbu are of course a type. You are right, it's +19dbu.
It works like this:
Vin for ADC full scale = 4V P-P for the CS5394 which is used in the Behringer EQ.
The transformer is 5:1 stepdown thus for full scale you need 4V * 5 = 20V P-P. Now 20V P-P = 7.07V RMS. And 7.07V RMS is appx. +19dbu.
Sayonara
Kuei Yang Wang said:
Ah, ok. Thanks.
Wasn't trying to be hypercritical or anything. Been working on a transformer-based power amp and have had dBu on the brain lately so that one just jumped off the screen at me.
Ah, ok. So the Behringer's being used between a preamp and power amp? Can't think of any source components beside that wacky DI/O that will give you an output on the order of 7 volts.
se
Koinichiwa Steve,
Well, it's simply that the "ironimous" approach was meant to be a 1:1 replacement for the Behringers analogue stage. As so much other Pro gear it is set to >+20db for full scale. This improves S/N Ratio and gives good headroom. And it makes sure the ADC's delta-sigma modulator is driven from a really low impedance. With the usual 200...5000 Ohm source impedance of anything preceeding the Digital EQ unit the drive impedance to the ADC is below around 25 Ohm under such conditions (this is important).
BTW, the 5361 the input voltage differs slightly, but I feel a stepdown transformer remains required to avoid impedance issues. Alternatively a very large capacitor is required across the ADC input (2.7nF) which would make the input a very difficult load.
Sayonara
Steve Eddy said:
Well, it's simply that the "ironimous" approach was meant to be a 1:1 replacement for the Behringers analogue stage. As so much other Pro gear it is set to >+20db for full scale. This improves S/N Ratio and gives good headroom. And it makes sure the ADC's delta-sigma modulator is driven from a really low impedance. With the usual 200...5000 Ohm source impedance of anything preceeding the Digital EQ unit the drive impedance to the ADC is below around 25 Ohm under such conditions (this is important).
BTW, the 5361 the input voltage differs slightly, but I feel a stepdown transformer remains required to avoid impedance issues. Alternatively a very large capacitor is required across the ADC input (2.7nF) which would make the input a very difficult load.
Sayonara
Kuei Yang Wang said:
Ok. So the Behringer's normally padding down its input expecting to see much higher input levels than would be expected from typical consumer gear, yes?
Yes. I can see the step down transformer being a plus for impedance issues, but it also rather precludes its use directly with a consumer source component or through the processor loop of a consumer preamp.
Now, if you wanted a 1:1 solution so you could drive it straight from a source component or through a preamp's processor loop, you could use a 1:1 output transformer. The impedance issue wouldn't be a problem (the winding resistance is so low it doesn't add appreciably).
The one caveat would be any coupling capacitance on the output of the source. You wouldn't be able to get away with the couple of uF or so that's often used. You'd need several hundred uF which pretty much means electrolytics.
A good compromise might be a 1:1 input transormer with a pair of simple FET buffers on the output.
se
Koinichiwa,
As is the case with all Pro Gear.
I cannot see why it would preclude anything. Normal sources have in effect "CD Standard" 2V RMS for what equals full scale. Now 2V is effectively 11db below "full scale". So in effect the top two bits are assigned to "permanent headroom". Given that the Behringer has relaistically 18 to 19bit analogue resolution end to end we are still well within what CD requires.
The winding resistance MAY not be an issue, the source impedance however is a major one. You will have to use the 2.7nF input capacitor to ensure that the peak currents caused by the delta sigma modulators sample/hold system can be accomodated.
This now limits source impedances drastically and provides a difficult load to drive. In the end you will have to buffer the input to ensure reliable impedance conditions. Where a 4:1 stepdown transformer with the CS5361 would make sure that up to 1600 Ohm source impedance can be accepted, using a 1:1 would limit you to no more than 200 Ohm, which can be more than a little issue.
Yes, there are many options.
I personally think simply using a stepdown transformer and cranking up the gain ahead of the AD is nevertheless the better solution as it maximises S/N ratio on a systemwide basis. Pro-Audio knows quite well what they are doing with these high levels.
On the positive side, the newly announced Behring DEQ2496 with 24bit/96Khz conversion and much better DSP and DSP features will have inputs and outputs that will be switchable +22/+12 dbu for full scale. This gives around 3V RMS for full scale, so the unit should be happy within any processor loop.
Sayonara
As is the case with all Pro Gear.
I cannot see why it would preclude anything. Normal sources have in effect "CD Standard" 2V RMS for what equals full scale. Now 2V is effectively 11db below "full scale". So in effect the top two bits are assigned to "permanent headroom". Given that the Behringer has relaistically 18 to 19bit analogue resolution end to end we are still well within what CD requires.
The winding resistance MAY not be an issue, the source impedance however is a major one. You will have to use the 2.7nF input capacitor to ensure that the peak currents caused by the delta sigma modulators sample/hold system can be accomodated.
This now limits source impedances drastically and provides a difficult load to drive. In the end you will have to buffer the input to ensure reliable impedance conditions. Where a 4:1 stepdown transformer with the CS5361 would make sure that up to 1600 Ohm source impedance can be accepted, using a 1:1 would limit you to no more than 200 Ohm, which can be more than a little issue.
Yes, there are many options.
I personally think simply using a stepdown transformer and cranking up the gain ahead of the AD is nevertheless the better solution as it maximises S/N ratio on a systemwide basis. Pro-Audio knows quite well what they are doing with these high levels.
On the positive side, the newly announced Behring DEQ2496 with 24bit/96Khz conversion and much better DSP and DSP features will have inputs and outputs that will be switchable +22/+12 dbu for full scale. This gives around 3V RMS for full scale, so the unit should be happy within any processor loop.
Sayonara
Kuei Yang Wang said:
Nevermind. I was thinking of the step-down transformer in addition to the padding of the Behringer's input stage, but you're using the transformer to replace it, yes?
Though I can think of a number of phono stages that haven't got outputs up in the 2 volt range.
Yes. Point being that if your source impedance is already low enough, using an output transformer wouldn't increase the source impedance appreciably compared to an input transformer which can add significantly due to greater DC resistance of its windings.
Yes. The step-down would be a big plus for more girly-man output stages. Most tube output stages would certainly choke on a load like that. But one might also argue to get rid of the girly-man output stage.
The high line levels certainly help with regard to noise picked upon on the lines during transmission. But does nothing with regard to self-noise, noise sources within the equipent itself, power supply noise, etc. And attenuating those signals isn't a noise-free process either.
So I'm not sure there's much to be gainded by that approach (i.e. cranking up the gain only to attenuate later) except where you have unusually long line lengths to deal with.
That sounds like a smart move.
se
Koinichiwa,
Yup, that's the idea, Transformer directly to ADC, no active circuitry whatsoever before the ADC.
Hmmm. that's a question of gain. Also note that LP's are commonly cut with up to 10db+ above the nominal 0db point at least in the lower midrange (LP has a dynamic range maximum at around 300Hz, which corresponds with the level distribution in real acoustic music).
True, but with a Outputtransformer (like for example the S&B TX-410LO or the equivalent Jensen unit) you suddenly no longer have any "universality" and are limited to extremel low Z sources.
Actually, noise sources depend, but with the noisefloor being mostly defined by thermal noise and various active noisesources increasing the Level usually improves S/N as the noisefloor is a given. In Mike Pre's this becomes a little more complex as the input noise becomes critical, but even there more level in the active circuitry USUALLY means better S/N in practice.
The attenuation of the Level with stepdown input transformers arguably is not noisefree, but the Johnson noise contributed by the Windings DCR is rather minimal.
Sayonara
Steve Eddy said:
Yup, that's the idea, Transformer directly to ADC, no active circuitry whatsoever before the ADC.
Steve Eddy said:
Hmmm. that's a question of gain. Also note that LP's are commonly cut with up to 10db+ above the nominal 0db point at least in the lower midrange (LP has a dynamic range maximum at around 300Hz, which corresponds with the level distribution in real acoustic music).
Steve Eddy said:
True, but with a Outputtransformer (like for example the S&B TX-410LO or the equivalent Jensen unit) you suddenly no longer have any "universality" and are limited to extremel low Z sources.
Steve Eddy said:
Actually, noise sources depend, but with the noisefloor being mostly defined by thermal noise and various active noisesources increasing the Level usually improves S/N as the noisefloor is a given. In Mike Pre's this becomes a little more complex as the input noise becomes critical, but even there more level in the active circuitry USUALLY means better S/N in practice.
The attenuation of the Level with stepdown input transformers arguably is not noisefree, but the Johnson noise contributed by the Windings DCR is rather minimal.
Sayonara
Kuei Yang Wang said:
Yup. And you know I'm a fan of transformers so no argument there.
Yes, it is a question of gain. Specifically the amount of gain the designer chose to use in their particular phono stage. And of course cartidge output isn't the same from one make/model to the next either, even within a given type (i.e. MC and MM).
You don't quite get universality either way. With a step-down, you have to rely on sufficient input level. With a 1:1 output transformer, you have to rely on sufficiently low output impedance.
And I'm not trying to argue that any one approach is inherently any better than another. Just that there are caveats for both and which will be the more ideal will depend on each particular situation.
Thermal noise and active noise sources such as shot noise are both level dependent so you don't get any improvement there that I can see. And the higher your output levels into a given load, the more current you have to source/sink. It may SEEM that you get an improvement in S/N if you're just comparing the signal to quiescent noise. But of course quiescent noise levels only apply under quiescent conditions (i.e. no signal).
So I still don't see anything better than breaking even except with regard to noise induced into the lines.
Certainly. But the point was simply that it it all adds up and I don't see that you come out ahead of the game. In other words, there is no free lunch.
se
going to use jt-11p-1 for now,
Howdy Koy,
going to use the jt-11p-1's i have on hand for now, I only need single ended inputs, and when the finances allow was thinking of the jt-11ssp-8ma, I am awaiting confirmation from jensen technical support.
The jt-11ssp-8ma will allow nice single ended or balanced operation, quite nice.
Mark
link1853
Howdy Koy,
going to use the jt-11p-1's i have on hand for now, I only need single ended inputs, and when the finances allow was thinking of the jt-11ssp-8ma, I am awaiting confirmation from jensen technical support.
The jt-11ssp-8ma will allow nice single ended or balanced operation, quite nice.
Mark
link1853
Re: going to use jt-11p-1 for now,
Howdy Koy, [/QUOTE]
Hey, link! Sorry I didn't put the name to the nick. Thought the name rang a bell.
Why the 11-SSP-8M? The 11P-1 will readily accept balanced or unbalanced inputs. Or were you wanting to switch between 1:1, 1:2 and 2:1 modes?
Don't think the 11P-1's going to be very suitable given its high output impedance (a little over 2k with a mere 50 ohm source). And loading could be a bit of a problem. Might be better to take Thorsten's approach and use something like the 10KB-D. But you'd still have to work out the loading issue.
Lemme know what Dale has to say.
se
Howdy Koy, [/QUOTE]
Hey, link! Sorry I didn't put the name to the nick. Thought the name rang a bell.
Why the 11-SSP-8M? The 11P-1 will readily accept balanced or unbalanced inputs. Or were you wanting to switch between 1:1, 1:2 and 2:1 modes?
Don't think the 11P-1's going to be very suitable given its high output impedance (a little over 2k with a mere 50 ohm source). And loading could be a bit of a problem. Might be better to take Thorsten's approach and use something like the 10KB-D. But you'd still have to work out the loading issue.
Lemme know what Dale has to say.
se
Re: going to use jt-11p-1 for now,
Koinichiwa,
I need to agree with Steve here.
Neither Transformer strikes me as usable without added buffers after the transformer. in order to keep the Delta Sigma modulators Sample & Hold circuit working in a linear fashion you need to feed the ADC inputs froma very low impedance, 200 Ohm or less.
Even 200 Ohm is too much really as this then requires a capacitor across the inputs (2.7nF nominally, can be adjusted slightly) to cope with teh impulse currents the S&H circuit draws.
If you want to eliminate active circuitry you will have to use a suitable transformer. a 4:1 or 5:1 Stepdown seems required, giving 8 or 10V RMS for digital full scale.
Sayonara
Koinichiwa,
Mark Hathaway said:
I need to agree with Steve here.
Neither Transformer strikes me as usable without added buffers after the transformer. in order to keep the Delta Sigma modulators Sample & Hold circuit working in a linear fashion you need to feed the ADC inputs froma very low impedance, 200 Ohm or less.
Even 200 Ohm is too much really as this then requires a capacitor across the inputs (2.7nF nominally, can be adjusted slightly) to cope with teh impulse currents the S&H circuit draws.
If you want to eliminate active circuitry you will have to use a suitable transformer. a 4:1 or 5:1 Stepdown seems required, giving 8 or 10V RMS for digital full scale.
Sayonara
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.