Turtlerock Forum

I am designing a preamp based on the INA217 by Texas Instruments



what I have done is construct it with the DC offset OPA137 in place, and no input protection diodes or DC blocking capacitors on the input, no phantom power at this stage.

I tried it without any of the DC offset stuff, and it sounded the same, as soon as I put electrolytic caps on the input 47fF and a cap for DC offset at the output 470uf I dont like the sound, now these are cheap caps and perhaps better caps would be sonically more pleasing.




Here is what it sounds like against two commercial preamps

https://dl.dropboxusercontent.com/u/13120658/Archive.zip

Everything is labelled in the zip...

Acoustic guitar track and 3 vocals

I set all the gains the same, and tried to be consistent with my singing.

Heres what I would like help with:

what do you think of the tone of it, at this point?

Adding input and output transformers, I am having trouble understanding the impedance implication of the two 2K2 resistors for the DC bias on the INA217 and as well what the transformer specs should be

I am going into a RME UCX which as I understand it has a input impedance of 10K.

Does the INA217 see the load of the mic at its input or the result of the 2K2 resistors?

What should the impedance rating/turns ration of the input transformer be?

Thanks for any help.

Hey if you know all the answers, and dont want to give it away for free, or its too much a pain in the ass to type it I am happy to pay someone for an hour of their time, and I could pick their brains over the phone,,,I would like you to be able to make creative suggestions towards the tone... Rob? Anyone?

thanks

Peter

I should add.

the noise floor on my preamp is -90dB fs.. which is very cool.

Also, the mic I used, is shall we say, not my favourite for vocals.

Its my Mojave MA200, as my Beesneez T1 is being repaired, for the third time, thats another story.

But I suppose, If I can make a preamp that makes the Mojave sound good.. then.. I am on the right track

Hi Peter

I'm far from an expert in these sort of things, but ....

(1) The INA217 has an inbuilt balanced line receiver, so you do not need an input transformer - in fact you would need to significantly alter the schematic to cope with a unbalanced input signal.

(2) You don't need an output cap if you have the DC servo circuitry already in place.

(3) You should be able to go with pretty much any 1:1 or 1:2 (for the extra few db) output transister. But ...

(4) Do stick in a small resister (20 to 100R) in series with the output.

(5) The INA217 does not "see" any impedance. Each side of the balanced input goes to a separate opamp inside the chip, configured as a variable gain voltage comparator, each of which is of 60MΩ (= practically infinite) impedance across its positive and negative sensing inputs. This means the chip can deal with a wide range of input impedances - that said, see (7) below.

(6) The 2 x 2k2 resistors provide a path to ground for the INA217's input bias current. A mic will see these as about 5K impedance, but really its about keeping the chip stable more than impedance matching.

(7) Do look at the datasheet comments on source impedances: under 10Ω the INA217 will oscillate, near that use the paralleled 47Ω resistor / 1.2uH choke network on each leg of the balanced input. Optimum source impedance is 200Ω, anything approaching or over 10K basically don't bother, use another chip.

(8) VRgain is ideally a rev log pot. Just checking you picked that up.

(9) Forgive me if you know this, but the OPA137 needs to be powered as well. Its not shown on the schema 'cause its assumed that everyone knows this. The output offset DC would be pretty small anyway, but this servo circuit would effectively eliminate it.

(10) Finally, my gut says that the path from the INA217 to the gain pot should be kept as short as possible, ideally mount the pot on the PCB near the INA217 itself.

Good luck, and may the forum gurus feel free to correct me if any of the above is wrong.

Chris

Thank you so much for replying, and in such a detailed and helpful fashion.

It is of great help.

Cheers

just to add.

I am running an input transformer in my bench design atm.

This is for two reasons, one, I actually like the sound of it. Also, its to avoid the issue of capacitors in the signal path for phantom power.


Also, Chris, why do I have to have a small resistor 20 to 100R on the output?

Atm, I am tossing up whether to leave the output unbalanced, or to incorporate a balanced line driver, I will I guess build the line driver and see what it sounds like, everything is driven by the sound for me on this.

I also see JLM make a Go Between Plus, which I am thinking of using for a couple of reasons, the biggest being it actually suits the exact Input transformer I am using and will save me having to mount it and takes care of phantom, phase reverse and pad.

I also was planning on designing and building a PCB to house the finished design, but I actually came across a PCB for sale on EBAY that was only 7 bucks and will suit, which is cheaper than I could have the board made for, I have ordered a couple and will give them a try.

I also, want to try an OUTPUT TRANNY and see how it sounds... and that would solve the balanced issue in one go as well.

thanks again for replying ...I thought I was talking to myself here on this.. 8)

cheers

Peter

Wiz wrote:I am running an input transformer in my bench design atm.

Don't get me wrong, an input tranny will work but you are doubling up on input buffers. If you have, and like the sound of the tranny, you don't need the INA 217 chip - you would be better off with a straight opamp (or better a dual opamp configured as 2 gain stages).

Wiz wrote:Also, Chris, why do I have to have a small resistor 20 to 100R on the output?

Provides a load for an output tranny to see, and promotes general stability of the output opamp. I have a vague memory reading reading something else along the line of cable capacitance issues, but don't ask me for details on that one. It will do no harm to the output signal voltage, and might do some good in the real world of cables and actual rather than ideal circuits.

I also see JLM make a Go Between Plus, which I am thinking of using for a couple of reasons, the biggest being it actually suits the exact Input transformer I am using and will save me having to mount it and takes care of phantom, phase reverse and pad.

I reckon I must have made up a dozen or more JLM Go Between's, they are the absolute Go To kit for adding functionality to a preamp design. I have a couple sitting in my current MILA project.

I also was planning on designing and building a PCB to house the finished design, but I actually came across a PCB for sale on EBAY that was only 7 bucks and will suit, which is cheaper than I could have the board made for, I have ordered a couple and will give them a try.

what, and miss out on all the fun? $7 is a pretty useful preamp PCB. Good luck with it.

Again Chris, thanks so much for replying.....


Will look into all your suggestions.


Cheers

Peter

Wiz wrote:I am designing a preamp based on the INA217 by Texas Instruments

ina217-large-1.gif

what I have done is construct it with the DC offset OPA137 in place, and no input protection diodes or DC blocking capacitors on the input, no phantom power at this stage.

I tried it without any of the DC offset stuff, and it sounded the same, as soon as I put electrolytic caps on the input 47fF and a cap for DC offset at the output 470uf I dont like the sound, now these are cheap caps and perhaps better caps would be sonically more pleasing.

Here is what it sounds like against two commercial preamps
Everything is labelled in the zip...
Acoustic guitar track and 3 vocals
I set all the gains the same, and tried to be consistent with my singing.
Heres what I would like help with:

what do you think of the tone of it, at this point?

Adding input and output transformers, I am having trouble understanding the impedance implication of the two 2K2 resistors for the DC bias on the INA217 and as well what the transformer specs should be

I am going into a RME UCX which as I understand it has a input impedance of 10K.

Does the INA217 see the load of the mic at its input or the result of the 2K2 resistors?

What should the impedance rating/turns ration of the input transformer be?

Thanks for any help.

Hey if you know all the answers, and dont want to give it away for free, or its too much a pain in the ass to type it I am happy to pay someone for an hour of their time, and I could pick their brains over the phone,,,I would like you to be able to make creative suggestions towards the tone... Rob? Anyone?

thanks

Peter

Hi Peter,

The INA217 is meant to be a pretty colorless transparent type all in one mic pre chip. The specs are
pretty good. I think their INA103 was a better chip but not sure if it's still available.

The OPA137 is used in what is called a DC offset correction servo. The 0.1u cap / 1M R will define
how slow it does it's correction. Servos are subject to various opinion but IME, it is in the signal path and
if you change this OPA it will affect the sound to some degree. The 137 is not very good for audio.
I recommend using a decent audio FET IP opamp for the servo. Something like an OPA134 or similar.

WRT noise the 217 has 1.3nV/rt Hz spec at 60dB of gain. This will equate to an EIN of about -128.5dBu
(with the usual 150 ohm source) which means it's pretty damn quiet if not at the absolute cutting edge.

WRT IP loads, 217 looking 'out' sees mainly the Mics OP impedance which is very low (a few hundred ohms)
The mic sees the 2 x 2k2 = 4k4 impedance.

Putting a transformer in front of one of these balanced IP instrumentation type amplifiers is not as stupid
as you might think. It gives a few advantages such as elimination of the Phantom blocking caps, true
balanced IP to the chip for pseudo balanced OP mics, good RF immunity, some nice sonic flavour etc. I seem to
remember working on a DDA console that had exactly that arrangement on the IP.

A suitable mic IP traffo for this amp will be a low ratio type. Usually 1:2, these are also the highest performance
due to the low ratio. The transformer will come with recommended loading / snubbing circuitry.

WRT OP transformer, also a great idea. If you want some mojo use a steel core, transparent use hi nickel.
Cinemag have some really excellent traffos. The best OP traffos are generally large multifilar types. Ideally
driven from a low impedance however you will need -some- resistor value at the 217's OP or it may become
unstable (oscillate). 50 ohms should be OK or you can try the Jensen OP load isolator of a 100 ohm resistor
with some wire wound around it. (check their website) This give the ideal 0 ohm drive to the traffo at audio
frequencies but 100 ohms OP Z for the 217 at RF for stability.

WRT OP coupling cap, you should be able to eliminate it and let the servo do it's job.

If you are going to use IP (or OP) electro caps everyone's got their favourites but I think Elna Cerafine
or Silmic are decent sounding caps. It's pretty subjective, have some fun trying a few.


Finally, don't skimp on the power supply and have some fun trying different PS bypass caps you may notice
they affect the flavour of the pre a bit.

Have fun, if you need any help pm me.

cheers

Terry

Terry

I can't tell you how grateful I am to you for taking the time to write such a informative reply.

It's of such help.

Thanks

Peter

Wiz wrote:Terry

I can't tell you how grateful I am to you for taking the time to write such a informative reply.

It's of such help.

Thanks

Peter

No worries.

Apologies for re iterating some stuff Chris P already stated, was pretty tired last night just skimmed thru
the thread. All his comments WRT stability etc are right on.

It's a good idea to get the design off the breadboard and use something like a solid copper clad pcb
(avail from jaycar etc) and do it 'dead cockroach' style.
The copper plane is much better for a ground and local PS bypassing of the INA.
Also these days this is important as chips get faster. It will probably sound better / be quieter too.

Also you can have fun adding some mojo to the sound with old components like carbon composition
resistors etc. As long as you don't put them in the very front end, the slight noise increase wont be an issue.

I'm actually working on a fully discrete, VHQ valve / SS hybrid mic pre myself.
When it's done I'd be interested for you to have a play and give some feedback.

cheers

Terry

zenelectro wrote:

Wiz wrote:Terry

I can't tell you how grateful I am to you for taking the time to write such a informative reply.

It's of such help.

Thanks

Peter

No worries.

Further to my last post, it's a good idea to get the design off the breadboard and use
something like a solid copper clad pcb (avail from jaycar etc) and do it 'dead cockroach' style.
The copper plane is much better for a ground and local PS bypassing of the INA. These days this
is important as chips get faster. It will also probably sound better / be quieter.

Also you can have fun adding some mojo to the sound with old components like carbon composition
resistors etc. As long as you don't put them in the very front end, the slight noise increase wont be an issue.

I'm actually working on a fully discrete, VHQ valve / SS hybrid mic pre myself.
When it's done I'd be interested for you to have a play and give some feedback.

cheers

Terry

would love to...

8)

Peter

Things are coming along great.

The little circuit board I got off EBay is going to work out, by using links I am able to use the pcb just as I want it. Seems to be a good quality so that should make it easier/cheaper than building it.

I have tried INA217 and the SSM2019 chips and I also got a pin compatible discrete version from Jim Williams and have been trying out all three. Big difference in price of those things, but, compared to the price of commercial preamps still way cheap.

I still want to try a THAT 1510.

Each of the three chips, all have a different sound. The Jim Williams one is great ..... Should be at the price. Cost me 70 bucks landed.

The transformers I got off Rick, two Billingtons and one Astor are great.

I still have to workout what case to put this stuff in. I would really love to do something different in that regard. I don't know what yet, but something either really old and Art Deco looking would be awesome, think Universal Audios old console 8).

I am only using 4 resistors and 4 caps on the board and atm they are carbon resistors and cheap ceramic caps both of which came from Jaycar. I would really like to try some top notch replacements and see how they compare.

That is what I am struggling with, locating which exact type to get... If someone could help me out with a direct link, I would be most greatful.

2K2 resistor
1M resistor
100R resistor
.1uF cap (power supply for chips bypass)
470pf cap ( for hanging of inputs of chip)

What are the top flight versions of those?

Cheers

Peter

An update..

It's all coming along well.

I have a few boards built and I am currently doing testing against my commercial preamps, on difeferent sources.

It's been more expensive than I imagines it would be, but I see that cost as self education.

I have been doing some testing , and have got it set up so I can record same performance into two different preamps at once, null test, abc test etc on exactly the same performance.

Interesting, maddening, and intheend I don't know if it will all end up pointless, or result I me parting with some of my commercial preamps.

I do have a couple of arrangements that are sonically really good.

Well its done.






Sounds really great.

Really enjoyed the process. Learned heaps. Step Drills, the greatest tool ever invented.....

This one is heading down to a mate in Victoria, I will build one for myself, and perhaps one other for another mate, and then I think I will hang the tools up... 8) for a while anyways...

Thanks to everyone for their help.

Hey Wiz,

where did you get your INA217's from? I'm doing a group build of these and they're a bit pricier than I thought they'd be.

Cheers

Text_Edifice wrote:Hey Wiz,

where did you get your INA217's from? I'm doing a group build of these and they're a bit pricier than I thought they'd be.

Cheers

Hey Dave

Hard to remember I might have to dig thrum emails tomorrow to find it, which I will.

I know I got a freebie sample. But I bought a few as well, and that 1510s and ssm2019s etc

How many you need? I will have a few laying about here.

Cheers

Wiz

Might as well update the thread for posterity whilst I am here.


Mate sent it back...wasn't to his liking.

I have it sitting on a shelf and don't really use it.

I have it fitted with two discrete op amps, made by jim Williams of audio upgrades.

It's very clean and quiet.

I just seem to gravitate to my other preamps.

Was fun, worthwhile and maintained my sanity whilst recovering from surgery.

I get it ou t wvery now and then and plug something into it


Cheers


Wiz

I only need a handful - I'll drop you a PM. I think it's useful as a learning project but maybe also as a battery powered pre.

Wiz wrote:Might as well update the thread for posterity whilst I am here.


Mate sent it back...wasn't to his liking.
I have it sitting on a shelf and don't really use it.
I have it fitted with two discrete op amps, made by jim Williams of audio upgrades.
It's very clean and quiet.

I just seem to gravitate to my other preamps.

That is the problem with 'clean' pre amps they are very often not musical / listenable.
There is a new slew of 'clean' type pre's that are making good ground such as Forssell (Fred)
NPNG, Gordon Instruments etc.

Personally I like the advantages of 'clean' type pre's such as resolution, stage etc but keep the
musicality. This is very tricky from a design perspective and using opamps / INA chips is not the way to go.

My latest pre has a discrete SS / tube hybrid single stage high gain core that is essentially very low distortion
(clean) but there are controls to add H2 and H3 components to tailor the sound. There's also a switchable
VHQ transformer in the OP which is always a good thing sound wise.

cheers