IC in place of DOA???

[khstudio]

I have some 2604 & opa627's & opa637's I'd like to try in my API Style pre's for fun & comparing:
http://www.groupdiy.com/index.php?topic=15010&highlight=api+yam

It's set up with the standard 6 pin DOA (990) footprint & currently running on 24 volt rails.

Can this be done without changing too much, like the feedback network (currently using 20k/47pf) or whatever?

I think the 2604 will run on 24v but I don't think the 637 or 627 will
but I've heard great things about them. & can lower my JLM ACDC Power supply down for that test.

I want to try this for fun & to see how my Yamaha DOA (NE80200) compares.
I've never heard a DOA before this one but would love to hear others like the JLM 99v, 990, Real API, etc...

This is the only comparison I have for now but I've read that these can rival "Some" DOA's if extreme headroom isn't an issue. (?)

Thanks in advance,
Kevin

 

[tk@halmi]

[quote author="khstudio"]Can this be done without changing too much, like the feedback network (currently using 20k/47pf) or whatever?[/quote]

There is no simple answer. You need to use a scope to see if the step response is acceptable at the output.

 

[Greg]

The major issue is loading. A typical IC cannot drive a transformer. Even some DOAs, like the Forssell, cannot drive a load below 600 ohms. I'm sure those Tamura transformers will smoke a 2604 or the like... POOF ! :sad:

 

[Family Hoof]

[quote author="Greg"]The major issue is loading. A typical IC cannot drive a transformer.[/quote]
Yes, loading is an issue. You need to add a follower stage with ample power dissipation after the IC. Maybe you'd ought to buy a JLM Audio Hybrid kit for this purpose.

[quote author="Greg"]Even some DOAs, like the Forssell, cannot drive a load below 600 ohms.[/quote]
While it's true that not all DOAs are created equal, I'm not so sure about your claim. I've driven 150ohms (600 reflected through a 1:2 xfmr) with a 992, and while it did get warm and the headroom turned to ****, no damage occurred. A colleague suspected that the JFETs IDSS would limit current draw and prevent the thing from burning up.

Also, there's always the copper resistance of an o/p transformer primary to consider, present even with no load connected, and this is much lower than 600ohms. That's the harsh reality of life without bridging and DRV chips. High current!

 

[khstudio]

I didn't think it was that easy... :sad:

What made me think of doing it is I saw (I think maybe it was Bauman/ the DIY Pill) where I saw where he tried a 2604's & 5532's with & without an OT in an API style pre.

Also, the JLM (Baby animal) can run on an IC - 2604 but I think the circuit compensates for it.

So if I were to try it, running unbalance (No OT, using a cap) into more modern 10k inputs would give it a better chance :?:

I gather it's not really worth the effort though... I'll probably just end up trying a few other DOA's in this circuit.

I've heard the character changes from one IC to another but this is my first DOA encounter (Yamaha NE80200). It's absolutely killer how RICH they sound compared to an IC... now I'm hooked & want more :green:

Thanks for your help,
Kevin

 

[khstudio]

While on the subject of output impedence:

I asked this before but...
My pre's OT is 600:600 & seems to marry well with the Yamaha DOA I have so what, if any, are the downsides of running it into a 10k input without any loading? (like a 620 ohm)
I did listen to it & it seemed to close in just a little but at the same time it seemed to balance the frequencies & tighten the LOW-End a little.

What would you guys do? :?

Kevin

 

[Family Hoof]

[quote author="khstudio"]So if I were to try it, running unbalance (No OT, using a cap) into more modern 10k inputs would give it a better chance :?:[/quote]
Not a problem. That's what chips are designed for; it's called impedance bridging, aka voltage transfer (as opposed to impedance matching, aka power transfer). In this system the output impedance is very low and the load its driving (the input of the next piece of gear) it at least 10 times this amount, such that the voltage drop across the load is at least 90% of the signal so there is minimal loss (remember greater R, with constant I, means greater V, because I times R = V ---> ohm's law). Modern transformless gear has electronically balanced inputs, which give an i/p Z of 10k or greater by defaut. It's just a characteristic of the topology. Also, not all transformer-balanced gear was or is 600ohm i/p. Some pieces have a 1:1 in front of a hi-Z transistor input, or maybe a stepdown xfmr, which would reflect a higher Z than what the secondary is loaded with.

[quote author="khstudio"]My pre's OT is 600:600 & seems to marry well with the Yamaha DOA I have so what, if any, are the downsides of running it into a 10k input without any loading? (like a 620 ohm)[/quote]
It may have little or no effect at all. Depends on the xfmr design. A transformer is a reactive (read: non-linear) device. It has inductance, DC resistance, and probably some capacitance as well. That's three out of three ingredients of a resonant circuit. Granted, a change in secondary load (and this is probaby not a purely resistive load, meaning it too is reactive) should, in theory, change the frequency response... as should the changing output Z of the op amp with various amounts of feedback appied, but that's splitting hairs, IMHO. Some o/p xfmrs are designed very well and freq response remains virtually the same between 600 and 10k. Others are not.

I did listen to it & it seemed to close in just a little but at the same time it seemed to balance the frequencies & tighten the LOW-End a little.

Well then... question answered. What more do you need to know?

What would you guys do?

Use my ******* ears and forget about this math ********! :wink:

 

[khstudio]

Quote:
I did listen to it & it seemed to close in just a little but at the same time it seemed to balance the frequencies & tighten the LOW-End a little.

Well then... question answered. What more do you need to know?

I just wondered if anyone else had the same experience, that's all.

Quote:
What would you guys do?

Use my ******* ears and forget about this math ********! Wink

I know I know :wink:

I said the same thing on the Nuendo forum years ago & someone mentioned - we wouldn't even have the technology to record without it.

I don't see anything wrong with wanting to learn how things are suppossed to be (Math wise) & then listen.


Thanks for helping me learn :thumb:
- that's why I'm here,
Kevin

 

[Handcrafted Tone]

[quote author="Family Hoof"]

What would you guys do?

Use my ******* ears and forget about this math ********! :wink:[/quote]

:grin:

I made this little board to replace the DOA in 312 and Jensen twin Servo circuits. I used the AD797's. Sounds really good.

I compared two identical 312's, one with the 797, another with the JH990C. I preferred the 797's on vocal because it had slightly little less high end, and handled sibilance better. It seemed the JH had a little more bandwidth, making it a better choice for stringed instruments.

It's definatly worth trying it for yourself.

 

[clintrubber]

The major issue is loading. A typical IC cannot drive a transformer.

It might not be the direction in which you want to go, but for completeness let's add 'stacking' opamps (putting a few in parallel).
Like for instance done here:
http://www-s.ti.com/sc/psheets/sboa031/sboa031.pdf

Regards,

Peter

 

[rodabod]

Thanks, Peter. That was good reading - I was never sure how to connect them in parallel.

Perhaps someone could try this arangement in a 312 circuit?

 

[clintrubber]

[quote author="rodabod"]Thanks, Peter. That was good reading - I was never sure how to connect them in parallel.[/quote]
It's often done in little mixers for the headphone driver - you can keep using the same type of opamp and then simply use two for the headphone driver.

And IIRIC the JLM DI-circuit does an alike thing for driving an input-TX.
Not sure about the exact topology; the resistors might have been arranged a little differently there.

Note the righthand side of R2: it senses the 'final' output, and not the output of opamp A1. So output voltage will indeed be 'as intended' by the R2/R1-ratio but headroom may suffer since the outputs of the opamps need of course to swing a little more to realize the 'requested' output voltage at the R3/R4-node.

Hey, it's actually TX: exchanging voltage swing for output current :wink:
Haven't thrown figures at it or any real thinking, so take all this with a grain of salt.

Imho adding a BD139/140 pair to an opamp is sexier (but requires a few more components).

Bye,

Peter

 

[Handcrafted Tone]

[quote author="clintrubber"]http://www-s.ti.com/sc/psheets/sboa031/sboa031.pdf[/quote]


[quote author="rodabod"]Perhaps someone could try this arangement in a 312 circuit?[/quote]

I'll order an OPA2604 and test it out in about a week. If anyone else wants to try it, here it is in 2520 footprint.

edit - changed V+ and V- from traces into plains

click here to download these pics as a PDF so they print the correct size.

R1 and R2 refer to the 51R resistors in the BB schematic above (R3 & R4 on that schematic) I added some optional power supply bypass caps. You can use .1uF and 1uF on each rail.

 

[rodabod]

[quote author="Handcrafted Tone"]
[quote author="rodabod"]Perhaps someone could try this arangement in a 312 circuit?[/quote]

I'll order an OPA2604 and test it out in about a week.[/quote]

:thumb:

 

[Gus]

two things

one look into thermal epoxy and heatsinks for the chip.

two think about your pcb layout find out what pin(s) might be usefull to conduct the heat from the chip and make the Cu area bigger for some heatsinking.

 

[Svart]

It's nice to see some innovation.

I've been thinking for some time of a design using paralleled gain stages.. adding or subtracting stages of course adds and subtracts gain, similar approach but for a very different reason.

 

[Gus]

I would think one would want to do something based off a 110 circuit first section. That circuit seems to have been designed to work with IC opamps. This has the feedback gain control and a divider after the transformer and before the opamp.

 

[Kit]

If opamp dissipation is an issue, you could "bridge" the primary for increased power output. Of course you need two opamps for this....

 

[Handcrafted Tone]

[quote author="Gus"]I would think one would want to do something based off a 110 circuit first section. [/quote]

Can you point me in that direction? What is "the 110 circuit" ?

Gus, I've edited the pics, is this what you meant by more Cu space? I only made the +/-V traces into plains, would any of the other pins generate heat?

[quote author="Kit"]If opamp dissipation is an issue, you could "bridge" the primary for increased power output[/quote]

Do you have an example of this?

Thanks for your input!

 

[rodabod]

[quote author="Handcrafted Tone"]I've edited the pics, is this what you meant by more Cu space? I only made the +/-V traces into plains, would any of the other pins generate heat?[/quote]

I think if the whole chip gets hot, then it might be an idea to make all tracks from the pads nice and wide to help sink heat from the chip.

I suppose it depends on how hot you think it'll get though....