Hi everyone.
On the appe 1, sometime the 22uf 25V are blue (pre-NTI) and sometime they are yellow (NTI).
On NTI boards, they can also be mixed...
But always in the same way... yellow ones for +5V and -5V and blue ones for -12V and 12V.
It seems to me that the last boards have just yellow caps...
Does someone know the difference between them?
Are the yellow ones better? Or can they support higher temperature?
Aurel
Hello aurel6814,
it was rather more the fact, which offer was cheaper ( at the moment the bunch for a lot of boards have been bought )....
the temperatire was not that problem for the capacitors - otherwise Woz would have supplied them with coolingsinks...
rather similar point is that - at "pre NTI" boards - decoupling capacitors ( 100 nF ) were of discceramic type and later NTIversion had more modern squaretyp of multilayer ceramic ( which did indeed filter away the spikes from powersource better ).....
sincerely speedyG
Interesting! Thanks for your answer
Quite frankly, this mix of blue and yellow caps is a bit of a puzzle to me. I think they are rated the same. One possibility, that accounts for the mix, is that they used pick and place automated machines to stuff the boards and different feeds were used for different locations. Another possibility is that different stations were used to manually insert different part locations, and when the blue ones were exhausted at each station they were replaced by gold ones. All conjecture - there probably is a better reason, one that we will likely never completely understand.
Regards,
Mike W.
I doubt pick and place machines were used. Not very common in 1976. I'm pretty sure the boards would have been hand assembled before wave soldering.
There's probably no significance in the mix of blue and yellow caps. I'm guessing that there were just two batches of capacitors that were mixed together and whatever came out of the box when the person assembling them reached in is what went on to the board.
Hello PhilPower,
i agree with the that as explained also above...
experienced guy ( and Woz was really experienced ) can solder 2 boards per day.
First batch of "pre-NTI" boards ( 100 pcs. ) could have been made within 2 months....
Second batch of 150 NTI-boards could have been made within 3 months after waiting 1 month
for the boards to be made from NTI. Not to forget at least 2 to 4 months time required to sell
the first batch. As we know - only 250 Apple 1 have been made all together....
so that explains the time from 1976 to 1977. And then Woz started to develop the Apple II Rev.0
which was made after another year of development time and in that time they were also hunting for
riskfinance and space for factory.
The use of machines in assembly and soldering started first with the Apple II !
sincerely speedyG
It's known that Woz didn't solder the boards (other than prototypes) by hand. The boards were assembled and wave soldered by a commercial PCB assembler. As I understand it Apple received the boards already soldered and assembled, other than populating the IC sockets. Woz and friends stuffed the boards with chips, and he would check each one and fix any problems with them.
The thing is the yellow caps are always in the same positions...
I don't think that's true. Looking at the pictures of NTI boards in the Apple 1 registry there can be anywhere from one to five yellow caps. There doesn't appear to be any pattern to it. It just looks like a random mix to me, which is entirely consistent with there being two batches of capacitors mixed in a component bin that was used for assembly.
It's pretty clear to me...
There is always one or three or five yellow caps and they are always on the same positions.
1 : there is a lot with just one on the +5V... that's why SpeedyG theory is interesting.
http://thedigitalstory.com/2012/10/28/apple_1_computer.jpg
http://www.willegal.net/appleii/images/a1registry/dave-arlg.jpg
http://www.digibarn.com/friends/jef-raskin/apple-1/DSC01712.JPG
http://www.willegal.net/appleii/images/a1registry/postman-fixedup-lg.jpg
3:
http://www.apple1.fr/images/apple1top.jpg
http://www.rickcrandall.net/images/32_Apple4.jpg
http://www.willegal.net/appleii/images/a1registry/37withbox.jpg
http://www.willegal.net/appleii/images/a1registry/comphistmusntilg.jpg
5:
http://siconic.com/computers/Apple/Apple-1/Apple1%20016.jpg
http://www.willegal.net/appleii/images/a1registry/910Apple1lg.jpg
http://4.bp.blogspot.com/_zxI22dsOsAo/S5B-X0UKw8I/AAAAAAAAAbU/URE_2ErXPnk/s1600-h/Picture+2.png
I hope I haven't post twice the same but there is lot's of exemple anyway.
There is one exeption but this one is pretty strange anyway...
http://www.willegal.net/appleii/images/a1registry/Larson3lg.jpg
I really can't see any pattern that can't be attributed to coincidence. The fact that there is an "exception" at all discounts any pattern.
If there were always, for example, three yellow capacitors in the same place then you might have a point.
There are different numbers of yellow capacitors on known NTI boards and they are not always in the same place. Their distribution on the PCB does not appear planned in any way.
In any case, the yellow and blue caps are the same value. What would be the point of putting them in particular positions?
Indeed! That's my question!
In my oponion, it can't be random :
If there is one yellow cap, it's always on the +5V
If there are three, they are always in the same spot. (+5V, -5V and cursor control)
We can see on pictures that yellow ones are from 76.
Maybe they just used more recent caps in the most criticals positions?
As explained in several postings related to the Apple-1 the power source is quite "noisy" - has quite bad spikes....
in the early versions of the Apple-1 ( pre-NTI ) rather more, than in the later NTI version -
this is in majority caused by the decoupling disccapacitors in preNTI version compared to the
sqared multilayer capacitors in the NTI version....
nevertheless even in the NTI-version there is quite a lot noise left on the powersourcelines
( and therefor on the clockcycles )....
on the other side there are not that many components getting power from the +12 V and - 12V.
Most components hang on the +5V and - 5 V ( like the shiftregisters and the amp that enhances the clockcycles
(DS0025) - so probably somebody might have had hope that the new caps filter better the spikes
- and that persons proabably set priority to popolate at least the 1 in the + 5 Volt branch and if more have been used ( 3 ) then also populate the - 5 Volt and the one close to the clockchip ( above the DS0025 ) too to get cleaner clockingcycles.
Anyhow that´s just a guess....
speedyG
We will never know!
But it seems a good guess to me.
I think you're reading *way* too much into the significance of the yellow capacitors. They are the same value as the blue ones. Almost certainly the reason there is a mix is that a couple of different batches of capacitors were supplied to make up the quantity required for the build.
What might explain the yellow capacitors appearing more often in certain locations is that a different parts bin was used for each component location on the PCB by the assemblers. The parts bins might have been filled originally with blue capacitors. When the supply in each bin was used up it was refilled with the new batch of yellow ones, similar to what Mike Willegal was suggesting for pick and place machines.
There is no electrical significance in having the capacitors in certain locations. In any case it is the -5V rail which is particularly poorly decoupled, not the +5V rail. The assemblers almost certainly would not have had any understanding about decoupling when placing the parts. If there were a design decision to use "special" capacitors for decoupling it would have been the same on every board.
This would only be a valid explanation if the Gold caps had much better ESR rating . Has anyone looked up the original manufacturer specs for both kinds of caps and compared?
regards,
Mike W.
PS I have an had an experiment in the works for sometime to see how much a little lower value ESR cap in the -5 volt circuit would affect noise on the -5 volt rail. Just haven't had time to run the experiment.
I've got NOS yellow caps and used blue caps from 1973. I'll test the ESR, but it's going to be difficult to analyse the result... (maybe they aged differently)
@Phil : maybe you're right...
or maybe there is a reason that we haven't found yet
I have tested the ESR...
blue caps : 1.5
yellow : 0.8
but... I don't know if I have exactly the caps you can find on an apple 1, and the blue ones have been used.