Op amp for MINT

[sgeyedoc]

Has anyone had any experiences using the AD8065 dual amp vs. the BB OPA2227 in the MINT? Any specifics on alterning any of the other compoents? Size issues for the board? Thanks.

 

[Megaptera]

I've never used either of these opamps or a mint board, but just looking at the datasheets, the 8065 has the same pinout, so you should be able to fit it on the board without a problem. According to Tangent's Opamp guide, it seems to work decently as a portable's opamp.

 

[tangent]

Quote:

AD8065 dual amp

The 8065 is the single. 8066 is the dual.

Quote:

Any specifics on alterning any of the other compoents?

You can drop an 8066 right into the circuit, no changes.

Quote:

Size issues for the board?

All SOIC-8 chips are the same size, by definition.

 

[Earwax]

The 8065 is the single and the 8066 is the dual version, isn't it? (yes, tangent answered this while I was typing)

I didn't have any luck with the 8066 in a cmoy or a47, it sounded rather thin and washed out. I haven't built a MINT yet, does this chip sound better in a MINT?

 

[tangent]

Quote:

it sounded rather thin and washed out.

It's a 145MHz chip. It's likely it was oscillating. The CMoy design isn't conducive to high speed chips, and my limited experience with the A47 (at least, the CHA47) says that one's even worse.

As for the MINT? Well, there's no extra bypassing relative to the CMoy, but the layout is very tight so the electrolytics are close by, there are fewer parasitics than in your average CMoy, and multiloop has a way of stabilizing marginal circuits. I wouldn't guarantee success, but I wouldn't be worried to try it.

 

[Earwax]

Quote:

Originally Posted by tangent It's a 145MHz chip. It's likely it was oscillating. The CMoy design isn't conducive to high speed chips, and my limited experience with the A47 (at least, the CHA47) says that one's even worse.

Yes, it probably was oscillating, but at the time I tried it I still didn't have a digital meter to check anything.

I haven't found the A47 to be any more or less stable than the cmoy, just different.

 

[Earwax]

Quote:

Originally Posted by tangent It's a 145MHz chip. It's likely it was oscillating. The CMoy design isn't conducive to high speed chips, and my limited experience with the A47 (at least, the CHA47) says that one's even worse. As for the MINT? Well, there's no extra bypassing relative to the CMoy, but the layout is very tight so the electrolytics are close by, there are fewer parasitics than in your average CMoy, and multiloop has a way of stabilizing marginal circuits. I wouldn't guarantee success, but I wouldn't be worried to try it.

I went back and tried it again and the 8066 sounded really nice. I started wondering why I'd disliked it so much before. After I bit of head scratching, I remembered, and tried it again, this time with very low imdepance phones, 26ohm E3s. The chip got hot enough to fry a pinkie on. Ouch. That's in a rather generic cmoy.

 

[tangent]

Would you be willing to repeat that test while you have an ammeter hooked up inline with the amp's power supply? I'd bet it shoots way up when you plug those low-Z phones in. By "way up" I mean a lot higher than you'd expect just from the impedance difference vs. whatever other phones you were using.

 

[Earwax]

Quote:

Originally Posted by tangent Would you be willing to repeat that test while you have an ammeter hooked up inline with the amp's power supply? I'd bet it shoots way up when you plug those low-Z phones in. By "way up" I mean a lot higher than you'd expect just from the impedance difference vs. whatever other phones you were using.

Oh yeah, I didn't think about it yesterday, but that much heat couldn't come from nowhere.

no load - a bit under 15ma
64 ohm koss phones - around 17ma
26 ohm E3s - 70 ma

An AD8620 can drive the E3s pulling just 7ma.
This is in a generic protoboard cmoy with a TLE rail splitter.

 

[tangent]

Gotta be oscillating. You could probably cure it with some nearby fast bypass caps.

 

[sgeyedoc]

Where would you put the caps? Values? Thanks,
Steve

 

[tangent]

Quote:

Where would you put the caps?

In a MINT, you'd have to solder-tack one leg of the cap to each op-amp supply pin. The other leg can wend its way lazily to ground.

In a less congested amp, there may be a better way. For instance, in an amp with DIP-8 op-amps, you could solder the cap to the underside of the board, directly to each supply pin.

Quote:

Values?

0.01uF or 0.1uF ceramic or film.

 

[jamont]

I have a MINT with similar problems. I noticed a whooping sound when turning up volume with no input. Square waves on the scope showed lots of ringing. I cured this by tacking a 10pf NP0 ceramic across the feedback resistors.

This amp was built with a OPA2228, I suppose that was a mistake. But I've used the DIP-8 version in a PIMETA with no problems, I thought it would work fine in the tighter layout of the MINT

 

[tangent]

If you have C4 installed in your PIMETA, jamont, that would explain the difference.

 

[jamont]

Quote:

Originally Posted by tangent If you have C4 installed in your PIMETA, jamont, that would explain the difference.

Yes, I do. But why does it make such a difference? Also, is this a better solution than mine?