Scratch Fairchild 660 Worth it?

[StarTrucker]

If you're going through the effort to do one, is it worth the extra few hundred bucks worth of transformers and tubes and just do a 670? Is a 660 something like 70% of the way to a 670 in terms of the amount of work it takes to build? It seems that way from the schemo, especially if you leave out the lat-vert switch.

For those of you who have done one or the other, what are your thoughts?

 

[pucho812]

I never was a 670 fan. I prefer duel 660’s vs a single 670.

 

[StarTrucker]

But a 660 is not half the work of a 670 is it? Isn't it more than half?

 

[beatnik]

I haven't made any yet but I am planning to.
I suggest you check DaveP 660 from scratch thread, lots of useful information.

The 670 has a single PSU for both channels.
2x 660 would need 2x PSUs.

This and the lat/vert switching would be the main differences I guess.

I think the extra cost is at least a couple of thousands than a few hundred particularly if you use Sowter transformers and JJ 6386 tubes.

Usually a tube stereo unit need to go through some extra tubes to select two matched sets that have same gain.

I have started watching a series of videos from Eric Valentine of UTA who makes the Unfairchild 670 recreations and gives some insight on how they make their units.
For example they made a lighter control amp by swapping the push pull pentodes with an ecc99 dual triode. I guess it allows for a cheaper output transformer and also saves a pair of tubes and their heater current.

 

[abbey road d enfer]

If you're going through the effort to do one, is it worth the extra few hundred bucks worth of transformers and tubes and just do a 670? Is a 660 something like 70% of the way to a 670 in terms of the amount of work it takes to build? It seems that way from the schemo, especially if you leave out the lat-vert switch.

For those of you who have done one or the other, what are your thoughts?

I have'nt done any, but I would say it depends very much on the intended use.
If you want to process stereo signals, I believe a 670 is essential, because of the MS capability.
Making two Vari-Mu compressors behave perfectly in sync is almost impossible, unless you have 100's of 6386's to select 8 perfectly paired.
Rein Narma was perfectly conscious of it when he designed the 670 and that's why he chose MS operation.

 

[abbey road d enfer]

For example they made a lighter control amp by swapping the push pull pentodes with an ecc99 dual triode. I guess it allows for a cheaper output transformer and also saves a pair of tubes and their heater current.

Which goes against the very essence of a 660. The power amp used to drive the side-chain is a distinctive point.
If you don't do that, you'd better build a 436.

 

[beatnik]

Making two Vari-Mu compressors behave perfectly in sync is almost impossible, unless you have 100's of 6386's to select 8 perfectly paired.

It seems the internal push pull balancing is not too hard to achieve, thanks to the 4 paralleled triodes on each side. So perhaps not so much of an issue for a 660.
However it maybe difficult to have two channels of the 670 match in level at the same time.
On the Unfairchild they added a trim control to fine tune the gain, altering the bias of the 6386 tubes.

One of the videos from the series I mentioned above is showing an interesting test, trying to make a set that is matched well enough out of 8x random JJ 6386 . For sure the gain trim control was useful to achieve level balance between the two channels which would have likely been impossible with just 8x tubes available.

I guess unmatched bias between the two channels would have the consequence of a different distortion profile, but I reckon that would be more difficult to hear than a level imbalance that messes up the stereo image.
Overall I think they made an intelligent choice, because it doesn't force them as the manufacturer or an end user in need of s retube having to buy so many 6386 - for which JJ has a monopoly and charging a fortune.

Which goes against the very essence of a 660. The power amp used to drive the side-chain is a distinctive point.
If you don't do that, you'd better build a 436.

I have the feeling they have chosen that route mainly to save costs or some kind of manufacturing compromise, but perhaps in practice the end result is not that different from the original ?
Basically they changed the tube in the 2nd gain stage (ECC81 instead of 12BH7) and the output stage (ECC99 dual triode instead of 2x 6973)

Personally I think even with the different control amp configuration it would be a bit unfair to compare it to a 436, having said that I would certainly be interested to know what you think the shortcomings of the alternative control amp configuration could be from a technical standpoint.

 

[abbey road d enfer]

It seems the internal push pull balancing is not too hard to achieve, thanks to the 4 paralleled triodes on each side. So perhaps not so much of an issue for a 660.
However it maybe difficult to have two channels of the 670 match in level at the same time.
On the Unfairchild they added a trim control to fine tune the gain, altering the bias of the 6386 tubes.

Ther is gain, and there is slope. In oreder to perfectly match two sets of 6386's, you need to match both their gains and their slopes. It's doable, but considerably reduces the yield.

One of the videos from the series I mentioned above is showing an interesting test, trying to make a set that is matched well enough out of 8x random JJ 6386 . For sure the gain trim control was useful to achieve level balance between the two channels which would have likely been impossible with just 8x tubes available.

I haven't found this video, but I'm not surprized by your conclusion, that making two channels work perfectly in sync is almost unachievable.

I have the feeling they have chosen that route mainly to save costs or some kind of manufacturing compromise,

Of course. Even with these savings, the final price is quite stiff.

but perhaps in practice the end result is not that different from the original ?

Probably, but "not that" different" is too different for some.

Basically they changed the tube in the 2nd gain stage (ECC81 instead of 12BH7) and the output stage (ECC99 dual triode instead of 2x 6973)

It's a considerable downgrade going from a 20W amp to a 6W. The power amp in teh side-chain is the distinctive element in the 660/670. It directly affects the capability of dealing with transients.

Personally I think even with the different control amp configuration it would be a bit unfair to compare it to a 436, having said that I would certainly be interested to know what you think the shortcomings of the alternative control amp configuration could be from a technical standpoint.

I was probably exaggerating, but it was to illustrate what happens when you start to add compromises. You may go from a 20W side-chain to less, you would go from 4 or 8 vari-mu tubes to only 2, you would ditch an intestage xfmr, and what do you get?

 

[Rob Flinn]

There are quite a few fairchild derivatives that use solid state amps for the sidechain. This makes the box significantly cheaper doing away with a bunch of valves and audio transformers. The EAR and the POM versions spring to mind. One could probably use a single chip amp quite effectively.

 

[beatnik]

Ther is gain, and there is slope. In oreder to perfectly match two sets of 6386's, you need to match both their gains and their slopes. It's doable, but considerably reduces the yield.

Right so that means the two channel might match ok at let's say 5dB compression but be off at say 15dB.
This is the video I was talking about

It's a considerable downgrade going from a 20W amp to a 6W. The power amp in teh side-chain is the distinctive element in the 660/670. It directly affects the capability of dealing with transients

The reason for such a powerful side chain is to have a faster attack without side effects like distortion or thumping or there is something else ?
An in depth analysis would be very interesting.

 

[beatnik]

There are quite a few fairchild derivatives that use solid state amps for the sidechain. This makes the box significantly cheaper doing away with a bunch of valves and audio transformers. The EAR and the POM versions spring to mind. One could probably use a single chip amp quite effectively.

Yes and the Stamchild too. Now look at the price of that and it makes the Unfairchild or the Herchild price look like a fair deal..
Plus the Stamchild is build like crap with wiring soldered directly to PCB without any kind of connector. I feel very sorry for who is gonna have to repair something like that...

 

[abbey road d enfer]

Right so that means the two channel might match ok at let's say 5dB compression but be off at say 15dB.

Predictably.

This is the video I was talking about

OK. It just shows that unbalance increases with the global amount of compression, which we all knew from the start.
Mastering engineers may not care because they use very little compression, but mix engineers may feel different.

The reason for such a powerful side chain is to have a faster attack

That and only that.
Rein Narma explained that he wanted to have the shortest possible attack time, because it's what protects transmitters and cutting heads.

without side effects like distortion or thumping or there is something else ?

Distortion and thumping are dealt with mainly in the vari-mu stage. It's the reason for the paralleled tubes.

 

[abbey road d enfer]

There are quite a few fairchild derivatives that use solid state amps for the sidechain.

That makes sense in the context of cost reduction. It may even bring new possibilities, such as providing even faster attack times.
However, people that are in the market for a 660/670 may put a limit on what's acceptable for them.
In a way, it's somewhat like installing a small block in a steam car.

 

[Rob Flinn]

That makes sense in the context of cost reduction. It may even bring new possibilities, such as providing even faster attack times.
However, people that are in the market for a 660/670 may put a limit on what's acceptable for them.
In a way, it's somewhat like installing a small block in a steam car.

Agreed. However the cost savings are not insignificant on what would still be an expensive build. I built a 670 and it took me 5 years to get all the parts together with 3 sets of NOS 6386's to future proof it. I bought the parts starting in 2001. I think, even then, buying the sowter transformer set, Davens & valves is was still £3-4k. With hindsight I wouldn't bother now, because of the headache involved & having to use the not so stable tesla 6386 etc even the 6973 in the sidechain aren't so cheap now. I would instead spend the money to improve my mic cupboard.

 

[abbey road d enfer]

I think building a 670 is a work of dedication. It's not going to make your next record a hit, or turn a lame song into a standard.
I'm not even convinced it attracts or retains customers, like a U47 or ELA251 would, so there's no financiary incentive.
It's something one should endeavour only if it doesn't impair one's other necessities.
I've given it a thought, as an intellectual challenge, but one is immediately stumbling on the incompressible cost of tubes and iron.

 

[pucho812]

But a 660 is not half the work of a 670 is it? Isn't it more than half?

Well the 660 and 670 share the same psu design. The 660 is often thought of as half a 670 but it’s not.
My reasoning for duel 660’s is because when I was starting out guys I worked with were kind enough to show me what is what and it is a world of difference when you compare a single 670 vs dual 660’s.
I happen to refer the dual 660’s then a single 670 on a stereo source.

 

[Ó Draighnáin]

I've thought about it but decided against it. I will never do it. You need time, a whole lot of precious time, it is going to take patience and time, to do it right.

I would also put time and resources into the mic cupboard (which is what I am currently doing).

 

[abbey road d enfer]

Well the 660 and 670 share the same psu design. The 660 is often thought of as half a 670 but it’s not.
My reasoning for duel 660’s is because when I was starting out guys I worked with were kind enough to show me what is what and it is a world of difference when you compare a single 670 vs dual 660’s.
I happen to refer the dual 660’s then a single 670 on a stereo source.

Well the main difference, except from the obvious absence of MS mode in the 660, is that it uses twice the number of vari-mu tubes in the input section. It's the only significant compromise Rein Darma made in it. Is it the reason you prefer it?

 

[pucho812]

Well the main difference, except from the obvious absence of MS mode in the 660, is that it uses twice the number of vari-mu tubes in the input section. It's the only significant compromise Rein Darma made in it. Is it the reason you prefer it?

As much as I hate to argue subjective, it’s always sounded better to me to use dual 660’s on a stereo bus then a single 670.

 

[abbey road d enfer]

As much as I hate to argue subjective, it’s always sounded better to me to use dual 660’s on a stereo bus then a single 670.

Since I never had the opportunity to do the comparison, I won't argue. I'm just trying to understand the possible objective explanation.