Date Coding for Hunts / Erie capacitors

[glowinganode]

Originally Posted by Farnze102

The new business was started on 21 May 1961, which happens to be week 21 of the year, and the preceding Sunday for that week was Whitsun day.

If this is the case, any capacitor using this code was manufactured after this date, discounting 1960

Originally Posted by Farnze102

So someone decided to use the codename WHITSUNDAY for dating capacitors - why? because it contained 10 none duplicated letters of the alphabet. Also the letters needed to have a corresponding number sequence so 0987654321 was used.

The sequence 1234567890 seems more logical (to me).
I'll have a look at the RA17 for a closer date.
It may be that the more common values (like .05uF) with a greater turnover would be more recent than the odd values that may have been sat in the stores for a while.
Rob.

[dinkydi]

Hi Rob, The sequence 1234567890 that you propose does not work with other data.

For example, it does not put Lucien's amplifiers in the correct order. Therefore, I am sticking with the original sequence for now.

Rob, can you help me by refining your data? You have 2 entries for YHW-1. Is this a duplicate entry, or 2 caps with the same value, or 2 caps with different values? Some of your entries show multiple numbers. This could indicate caps with the same value, or caps with different values, or a mixture. These possibilities have to be handled differently statistically. Could you give us a list which does not include duplicate caps (with the same value and same code). That is, if you have 5 x 1uF caps all with the code THW, only put THW-1 on your list. If there is also 1 x 6uF with the same code THW, put THW-2 on your list. I hope this is understandable.

Peter

[glowinganode]

Ok Peter, no problem at all.
Indeed YHW is a duplicate on the list, there were two of these.
Statistics was never my thing, thanks for the pointers, I'll refine the list.
I'll also look at electrolytics too, didn't occur to me before.
Rob.

[Farnze102]

Just to help things along. Please use the attached spreadsheet and map any the capacitor date codes. My guess is that a pattern will form indicating what code was used and when.
Ed.

[glowinganode]

Ok Peter, I've broken down the codes for you. Be aware though that the RA17 is built up from sub chassis's which may have been manufactured a few months apart. Unfortunately the caps are mixed up, so I can't say which cap came from which chassis.

W49B500K 0.1uF 150Vdc
HIW - 5
HYW - 4
WSW - 2

W49B501 0.25uF 150Vdc
WSW-1

W49B503K 1.0uF 150Vdc
WAY - 1

W49B511K 0.05uF 350Vdc
WSY - 5
YIW - 1
WSW - 8
TWY - 1
HIW - 1
WHY - 1

W49B512K 0.1uF 350Vdc
YHW - 2
WYW - 1

W49B521 .025uF 350Vdc
WSW - 1

From a spare VFO, definately newer than mine as it uses newer brown moulded resistor rather than the older ceramic "rock-n-roll" ones.

B500
TSW - 1

B501
IHW - 1

B511
HYW - 2
HIW - 1

B512
WSY - 1
YHW - 1

From a scrap chassis, about the same age as VFO;

B501
WYW - 1

B502
IHW - 1

B503
WAY - 1

B512
YHW - 1

Hope this helps,
Rob.

[Lucien Nunes]

Just looked in a couple of Hunts catalogues and found that a few of the pictures are large enough for the print on the caps to be legible. Of course they have usually arranged things so that the Hunts logo is visible and the date code hidden, or else they are so heavily retouched that the code has disappeared. But a few are fairly clear:

Catalogue C262, October 1948:

Various electrolytics all WYU
Various metalclad papers all WWD
Various metallised papers with AN where the 3-letter code is normally printed
Two values of silvered mica WTU

Catalogue C269, August 1955:

Only two types legible, both TTD

I have received more amplifiers with likely dates, will have another go at listing their caps.

Lucien

[GrimJosef]

It's taken a while but I think I've cracked this.

For those of you who just want the answer the code is WHITSUNDAY=1234567890 which is actually the simplest possible choice. And the three letters do indeed represent week (two digits) and year (one digit - the last one e.g. the 4 in 1954). There is no offset in the year so, for example, 1954 (and 1964) is represented by the letter T. There is an irritating wrinkle however. It seems that there were two possible orders of the characters. Some (most ?) capacitors were labelled using the Wk Wk Yr scheme which has become the consensus choice in the later stages of this thread. Others were labelled using Yr Wk Wk, as was suggested earlier. I don't know of a sure-fire way of working out, for any particular capacitor, which scheme was used except, obviously, that the week code must lie within the range 01 to 52. Occasionally this condition would be satisfied by either scheme though .

The key I used to work the code out was the Quad II amplifier. There are three independent ways of dating these, two of which are pretty well understood. So, like the Rosetta stone, the Quad II can be used to decipher the third one - the Hunts code.

Each Quad II has a serial number. Despite rumours to the contrary (mostly put about by people who are trying to sell two widely spaced amps as 'a close-matched stereo pair' ) it appears that Quad did use these sequentially, at least down to the level of a few weeks' production. Not only does the serial number let us put amps in order, but there is also a commercially available list which matches serial number ranges with individual years. The list seems to be not quite perfect, but overall it is pretty good.

As well as the serial number the Quad II contains a double electrolytic 'smoothing block' (labelled C4/6 on Quad's official circuit diagram). This can be dated using the TCC code which was discussed here https://www.vintage-radio.net/forum/...ad.php?t=71659. This gives the Year (starting with A in 1945 and going to Z in 1968, missing out I and O) and the month (A to M, missing out I). The only problem with this code, and to a lesser extent the serial numbers, is that in the years since the Quad II went out of production quite a few of these smoothing blocks have failed and been replaced with ones cannibalised from other amps. The only way to resolve this is to build up a large enough sample of amps and thereby to identify the replacements statistically.

As well as the serial plate and the smoothing block each Quad II originally contained four Hunts caps - one Mouldseal (C1), two paper-in-oils (C2, C3) and a small electrolytic (C5) as shown in the first picture. I own a number of Quad IIs and I know other people who do too. I've managed over a little while to accumulate data from 26 amps spanning 14 years of production and containing, between them, 80 Hunts caps. There isn't space for me to lay out all this data here, and some people would not want the serial numbers of their amps posted on a public forum. But by way of a check I've shared all of the data privately with Lucien Nunes and he's happy that it is not seriously flawed. What it shows is that with a very few exceptions the Wk Wk Yr code dates the Hunts caps exactly where we would expect to find them i.e. from the production year of the amp, according to the serial number, or from the year before. In addition there is also very good general agreement with the TCC code dates.

As a small example here are the codes from four amps with serial numbers between 30,000 and 33,000, covering roughly 4 months' production in late 1960 and early 1961, according to the serial number. The codes are in the order C1, C2, C3, C5 C4/6

HSY SYA SYA IDH RD
*** WYY WYY *** RJ
SYA WYY SYA HNY RG
HHY ITY IHY TAY RK

Converting to months these codes correspond to

Jun-60 Dec-59 Dec-59 Sep-62 Apr-60
*** Mar-60 Mar-60 *** Sep-60
Dec-59 Mar-60 Dec 59 Jul-60 Jul-60
May-60 Aug-60 Aug-60 Dec-60 Oct-60

Among the 14 Hunts codes here all but one lie where we would expect in terms of the amps' production dates. The exception is the Sep-62 electrolytic in the first amp. In fact in the whole 80-cap sample there were just three anomalies like this. Inspection of the components shows that the other two were clearly due to replacement caps put into earlier amps. But this electrolytic doesn't look immediately like a replacement. I'm afraid that for now it is a puzzle. In addition to the replaced Hunts there were also just two obvious replacement smoothing blocks out of the sample of 26. And in one amp all the cap dates were relatively close to one another but the serial plate date seemed too early. In this case I strongly suspect that the serial plate has been swapped.

The final clincher for me is the week codes. The sample of 80 caps generated 36 different week codes between 01 and 52 and not a single one outside this range. I haven't done the statistical analysis but it seems overwhelmingly unlikely that this could have happened by chance. So personally I think that for the Hunts caps in the Quad II amplifier the date code is cracked.

The same scheme also explains many of the codes reported by others in this thread. Again I haven't got space to go through them all but for example all of the caps in Rob's RA17 date from 1960 or the first half of 1961 - perfectly consistent with his 1961 build date. The caps in Lucien's Amp1 and Amp2 all date from 1963-65 or, as he put it, the mid-60's. And every single one of these, along with David's 'bag of 5' and Lucien's catalogue caps, give valid week numbers.

The wrinkle arises when we consider Richard's (Mr Moose's) electrolytics (all but the 8+8uFs) and the caps in Lucien's Amp3. Using Wk Wk Yr almost all of these generate invalid week numbers . Given the great results from the Quad IIs I was disappointed when I saw this. However I noted both Richard's and Lucien's comments that some of their codes were clearly gapped between characters 1 and 2. Lucien has since confirmed to me that in the case of his Amp3 this applies just to the 'troublesome' caps coded UYA and UWW. So I tried the Yr Wk Wk scheme on these and the dates now came good. Lucien's are week 9 and week 11 1966, a couple of years after his Amp1 and Amp2 data. Richard's now all date from 1957/8 and all have valid week numbers. We might imagine that the gapping would be sufficient to identify the Yr Wk Wk caps. But this isn't obviously the case. The second picture shows a couple of Quad II caps. They appear to have gaps between characters 1 and 2 but for both these codes (TAT and TIT) the Yr Wk Wk scheme would have generated invalid week numbers. They must, like all the other Quad II Hunts caps, be Wk Wk Yr. So perhaps all we can say is that if there is a gap between characters 1 and 2 then the Yr Wk Wk scheme may have been used.

Cheers,

GJ

[GrimJosef]

Ooops ! In my rush to finish the above post I made a mental miscalculation . The TIT code referred to right at the end generates a week number of 34 using the Yr Wk Wk scheme which is not invalid of course.

However a quick rummage has turned up another Quad II Mouldseal with a gapped code (see pic). In this case the HDU code would correspond to week 86 using Yr Wk Wk and that one is invalid.

Cheers,

GJ

[glowinganode]

Good effort GJ, well done.
Rob.

[dinkydi]

First prize goes to GJ. Many thanks GJ for your sterling work and persistence in cracking the code. Congratulations also to Rob who first suggested the "logical" code sequence. Thanks also to the data collectors and Farnze102 who set us all off - the word "WHITSUNDAY" will never be the same!

Peter

[GrimJosef]

Thanks to everyone else here too . I neither would nor could have got into this without the WHITSUNDAY and 2-digits-week/1-digit-year info. And I was just lucky with having access to a lot of data. The result will be very handy for me too. There are amps out there with no serial plates and the vulnerability of the smoothing block makes dating using just that suspect. The Hunts caps are sufficiently tricky to change that if there are any present then they are likely to be the originals and a very good guide .

Cheers,

GJ

[Lucien Nunes]

Yes it's a grand piece of detective work on GJ's part, with every exception analysed and understood. Those aside, the deviation of individual cap dates from likely amp dates was surprisingly uniform, any other likely sequence for the cipher invariably producing much wider deviation if not actual invalid results.

I've sorted out some of my data and uncovered a large stash of gapped codes on various types of cap from 1964-5, many within two amps of very similar serial number from around 1965 as corroborated by the parts with ungapped codes on the same chassis. They include:
Electrolytics W YS, T DT*, H AS*, W DS*, I TS.
Moldseals T IT, T WT, H AT*, Y UT*, H TT, W TT.
All of these decode to a likely date when read week-week-year despite the gap. Those with asterisks would yield an invalid week number if read year-week-week.

The only gapped codes of mine so far that seem to need year-week-week decoding are the ones GJ referred to earlier, U YA and U WW. Amongst these and Richard's gapped codes, the first digits are all 6, 7 or 8. Although there is insufficient evidence so far, it seems that the gap is irrelevant in caps made before 1966, but starts to indicate year-week-week from at least week 9 1966 onwards. Does anyone have any ungapped codes starting with U, N or D?

New evidence suggests the serial number of my original amp No.2 was wrongly quoted, I was reading it off the pack of parts removed that contained the dated caps. However the writing had smudged and rechecking the amp chassis itself reveals that it is lower than amp 1, which tallies with the cap dates. It is also possible, although no longer necessary to assume, that its number comes from an independent sequence.

The next step will be to test the code on Erie components. And of course start work on Dubilier...

[Heatercathodeshort]

I think the INIGMA machine would have struggled to work that one out. Well done! J.

[Neil Purling]

When I encounter Hunts capacitors they usually wind up in the bin. I don't bother looking at the date code......
I am somewhat surprised that Mouldseals were used in Quad amplifiers. I assumed those big blocks were Paper-In-Oil capacitors, not an assembly within, you learn something new every day.

[Lucien Nunes]

Hunts capacitors .... usually wind up in the bin

Fair enough, but now if you look at the date code in the process you can discover the date of the set even if no other means of dating is available.

As I mentioned early in this thread, one of my applications for the Hunts date codes is to date electric organs. Some of the larger instruments were one-off commissions for wealthy individuals, costing maybe £5,000 in the 1960s - the price of two average houses. For many of these specials, exactly zero documentation exists. No sales brochures, no owner's manuals, no factory records, not one man alive who remembers even what model name it was given. This is where you need every possible clue from the machine itself if you want to trace its history and provenance. And I like playing the detective.

Lucien

[GrimJosef]

I agree Lucien. Sometimes information about a set is considered unimportant. But once it's gone, it's gone. The capacitor codes are going to be a handy way of knowing whether an amp's original or has been 'got at'.

Cheers,

GJ

[Refugee]

I will have to get the numbers of the caps in the amp that I am fixing.
It is a Bell and Howell and there are pictures of the output transformer in the success stories. It has Hunts caps and I have been looking for something with a date on it.

[Radio Wrangler]

One thing to bear in mind is that the code, whatever it was, had to be simple enough that the people at Hunts could use it

On the serial numbers of classic HP test equipment, the first one/two digits are years since 1960, then 2 digits of the week number in the year, then a letter for country of manufacture A= America, U=UK, G=Germany. The remaining digits are a chronologically issued individual serial number. Way back then, serials were started at 100. Nowadays with competitors trying to analyse each other's serials to try to determine sales rates, I wouldn't be surprised if non-chronological individual numbers have become the norm.

If Hunts had only known the reputation they were to gain, I don't suppose they'd want to print any incriminating evidence on the things. Still they did feed and clothe a generation of TV/radio repairmen's families. (I still feel gratitude for bad drivers, they paid for my childhood)

David

[dinkydi]

Hi David, Just to be clear, the first 4 digits of serial numbers on HP test equipment correspond to a date, as you describe, but that date is not the date of manufacture of the equipment. Instead, the date is when the last manufacturing/design change was made - the date is a revision/modification number, if you like.

The HP manufacturing date is not given in the serial number but is equal to, or more recent than, the date in the serial number - manufacture could be 5 or 10 years later, if there was no manufacturing/design change.

Peter

[Refugee]

I am looking for a date on this old Bell and howell.
The mains capacitors are indeed Hunts however there are two reasons I would not like to trust the dates.
Firstly they were not that reliable and secondly the estimated date of manufacture falls in a 10 year period and you guessed it the code is at one end of the range.
The one cap I have replaced with a replacement from a later retrofit part has the code YDT with no spaces and is covered by the clip in service and is very difficult to see due to the mounting location.
It is 32uF res + 32uF smoothing 450V and on the same line it also shows the code:- K B 564.
The alternative way of dating this amp is the Dubilier capacitors showing the code KB.
They are Type 4700B 0.1uF 500V. I also have another amp with Dubilier code JH. All these 0.1s had the same code in each amp.
The other ways of getting a rough date is what they were used for.
They came from 16mm cinema amplifiers that were made for optical sound track (the ones that blew a raspberry if the film slipped off the sprocket) so new build of this equipment would be 1944 to 1954 based on them having octal valves as there were later ones with all glass 9-pin valves that would have been made up to about 1964. There were a few retrofits to fit magnetic sound and the replacement cap came off the PSU from one of these.
There are some photos in Success Stories.