On Friday, August 6, 2021 at 3:25:43 PM UTC-4, Joerg wrote:
> On 8/6/21 10:24 AM, Rick C wrote:
> > On Wednesday, August 4, 2021 at 3:30:32 PM UTC-4, Joerg wrote:
> >>>
> >> This is where legislators have to get people with tech know-how
> >> involved, for once. Example:
> >>
> >> I lost two Dell PCs in short order and one of them was a pricey
> >> i7-based power machine. Motherboard failure. Well, with the i7 it
> >> could also be a CPU failure but because no schematics are published
> >> it's not possible to tell. Short of shelling out lots of money to
> >> try a new CPU and possibly discovering that this was a waste of
> >> money.
> >
> > Really? You are going to debug a 1100 pin CPU??? Ok...
> >
> No. Finding out via the schematic which pin connects to a switcher
> circuitry to enable a start is all that's needed. Without a schematic
> you can only try whether a CPU swap makes it work again.

You can't trace the switcher schematic to see where the enable comes from? That seems a much simpler job. The last think I would suspect is the CPU..

> >> So, if PC manufacturers just published the schematics many
> >> computers could be kept out of landfills. Motherboards usually
> >> become unobtainable as spare parts after a short number of years
> >> and besides, throwing away a whole motherboard just because a
> >> little IC or a diode on there has failed doesn't make ecological
> >> sense.
> >
> > There is also a need for someone who can repair such devices
> > economically. With many devices costing only a few hundred dollars
> > it doesn't make much sense to make it repairable when the shop will
> > charge $200 or more.
> >
> If a device costs $1k and a schematic eanbles you (or a shop) to repair
> it in under 30mins it does make sense.

I've not seen a repair shop that would even diagnose a unit for under $100. Unless you are talking about someone essentially doing this as a hobby a repair shop has to charge rates similar to auto repair, a plumber or whoever. None of this is cheap. What type of unit is worth $1,000 when it breaks? It may have cost $1,000 new, but I'm not going to spend a couple hundred dollars on a four year old computer. Computers become obsolete more than wear out. A small minority of users will waste their time trying to nurse old programs to run on a new machine or try to keep an old machine running when newer software won't run on it correctly. XP is a perfect example. I don't waste my time trying to run the same old OS on every computer I upgrade to. I just get on with my work.

> We also need to think about ecological consequences. Example: Bought a
> clock radio for aruond $10. Came home, plugged in, several segmetns in
> the LED display were dead. Hurumph! Opened it up, spent about 20mins
> repairing some cold solder joints, done. It still works after more than
> 20 years. The alternative would have been this:
>
> Driving to the store, 30mins. Standing in line for the return, 5mins.
> Drive back home, 30mins. What is better, 65mins or 20mins? Oh, and then
> there is the matter of a gallon of gasoline that I avoided burning.
> Besides, this radio would have landed in the trash and I do not like to
> unneccesarily pollute our environment or that of others. We all have to
> do our part.

So you suggest the typical user should get a bad unit, attend Devry training school for some months to learn how to repair it and then attempt to apply their new found knowledge to repairing the $10 clock they bought half a year ago?

BTW, only a complete idiot would make a special trip to the store to return anything. So the hour and the gallon of gas are bogus. As to your pollution, you are already a blight on the earth just like the rest of us. Do you apply the same consideration to pollution issues when you take a vacation or visit family out of town or even just keep your home warm or cool? A clock radio isn't going to tip the scales.

> >> I would not expect any company to divulge trade secrets such as
> >> the contents of an FPGA. However, I do expect release of schematics
> >> like was standard practice in the 60's and 70's where,
> >> consequently, I was able to repair a lot of TV sets as a kid. There
> >> is no secret in schematics since any competitor could easily
> >> reverse engineer boards to derive the schematics.
> >
> > If there was actually a market for schematics someone would start
> > reverse engineering popular devices and selling them. It's not a
> > copyright violation to reverse engineer such devices.
> >
> Why is it that _all_ my ham rdaio gear came with schematics and service
> manuals?

I can't believe you don't understand that issue. Literally by definition ham radio operators are a crowd that understand how their products work, at least for the most part. Clearly the design info is important to that market. Why would you not understand the difference with a clock radio?

> Same with radio and TV set up into the 70's. My wife's big cassic Saba
> 8100 stereo from the 70's developed an issue. Thanks to the availabilty
> of a full serice manual with schematics I was able to quickly diagnose
> and repair it. FM reception was restored and we still use it every day
> in the living room.

Yeah, because up to the 70s people could and would repair such gear as it was relatively expensive. Now such gear is much cheaper on a relative scale and much more reliable as well as MUCH more complex, so harder to repair.

> > FPGA firmware and CPU software are another matter. Companies make
> > many repair actions a crime by enforcing copyright on things that
> > have any such code. You might think that if you own a board with
> > such code you could buy an unprogrammed board and copy the
> > copyrighted materials, but no, you don't have that right. That is
> > one of the issues the Right to Repair organization is addressing and
> > has won support in Congress for.
> >
> Well, let's hope. WRT computers I have left that era behind me when
> Windows 7 was called off. They are all on Linux now.

They still have proprietary firmware for booting up.

> >> Easy. Make that reportable and when it festers some goons would be
> >> sent out to that company for a little audit.
> >>
> >> It has to be reasonable though. For example, an ASIC with 100 pins
> >> or more or a BGA may not be considered serviceable by ordinary
> >> repair personnel. This is all more about common sense stuff. Such
> >> as when the switch of a old vaccum fries. That happened to us
> >> earlier this year. Could not get the spare, had to buy a whole new
> >> vacuum.
> >>
> >> I think this is one of the few laws that makes sense if
> >> implemented reasonably and without undue burned on manufacturers.
> >> Such as "Identify the five most common easy to fix failures on your
> >> product and make those parts available at reasonable cost".
> >> Switches, motor brushes, cables, gaskets, that sort of thing. For
> >> electronics it's schematics, parts can generally be bought at
> >> Digikey. I don't want a spare for a huge FPGA to be avaliable for
> >> three decades, I just want to be able to have a "map" that allows
> >> me to find that D17 is bad.
> >
> > Sure, but very, very few would benefit much from such laws because so
> > few can effect a repair even with that info.
> >
> Not really. It would open up chances for new business, the neighborhood
> repair guy.

You mean the guys who don't exist?

> You wouldn't believe how many people I met who open and
> close their garage doors by hand because the electric opener has gone
> bad. "I don't have that kind of money". Often it's just a corroded DIP
> switch, relay or the RF transistor has gone bad after a thunderstorm
> years ago.
>
> I repaired the (back then painfully expensive) Grundig Satellite allband
> radio of my geography teacher. The TV shop had given up but it was easy.
> One of the IF coils had no conductivity. Took the can off, copper had
> oxydized away. New winding, align it, done. And I was just a school kid.
>
> Without a schematic this would have been much tougher.

Yeah, YOU repaired such gear. 99.99% of people can't and don't. Why can't you see that? You always show a very limited view of the world from your own little corner without understanding how anyone else views it.

--

Rick C.

-- Get 1,000 miles of free Supercharging
-- Tesla referral code - https://ts.la/richard11209

On Sunday, August 8, 2021 at 4:47:22 AM UTC-7, Chris Jones wrote:
> On 08/08/2021 08:23, jla...@highlandsniptechnology.com wrote:

> > There's lots of sourdough around that started as flour, water, and
> > whatever drifted in through the window.
> >
> > We met the Truckee Sourdough lady and she says that's how hers got
> > started, in the back room of her husband's failing deli.

> I can't understand the popularity of sourdough - the selling point seems
> to be that it uses some random microbes that happened to be present in
> the kitchen, instead of ones that have been chosen by bakers over the
> centuries because they make nice bread. Like instead of eating normal
> vegetables why not go and pick some random weeds from a back alley near
> the restaurant, that is bound to taste better and make you healtier.

The famed French bread is sourdough (made with a starter); the pure yeasts
that make good (beer, bread, etc) are monocultures, but sourdough is
not, having a variety of microorganisms that work together in harmony.
Flavor of a sourdough (or lambic) culture, not species exclusivity,
is its value.

Weeds can be good food; dandelions are a salad green, for instance. We're
descended from gatherers, you know.

On Friday, August 6, 2021 at 3:51:33 PM UTC-4, Don Y wrote:
>
> We have firms that buy boards just to lift the gold plate.
> If some FR4 gets discarded... <shrug>

People make a big deal of windmill scrap. I wonder which is a greater issue, PWBs or windmills? We sell something like 2 billion cell phones each year... just cell phones! I bet the cell phones generate a lot more scrap than the share of the windmill scrap that was used to charge cell phones.

--

Rick C.

-+ Get 1,000 miles of free Supercharging
-+ Tesla referral code - https://ts.la/richard11209

On Sun, 8 Aug 2021 11:28:43 -0700, Don Y <blockedofcourse@foo.invalid>
wrote:

>On 8/8/2021 11:05 AM, Jeff Liebermann wrote:

>No, that's a naive expectation. You don't need to keep an entire
>page in RAM to be able to generate a page *in* the computer.
>
>You treat the page as a series of smaller regions (the top 1", the
>next 1", etc.) and define a clipping region into which you "draw"
>the page.
>Then, ship that portion off to the printer.
>Then, move the clipping region to the next region and repeat the process.

That works nicely for printers that can tolerate start/stop paper feed
operation, such as inkjet printers and plotters. Laser printers need
an uninterrupted paper feed (to keep the fuser from setting fire to
the page when the paper stops) which is not likely to happen if the
computah has to send parts of a page to the printer. Even with some
buffering, any interruption in data transfer will cause a misprint
simply because the paper cannot stop moving in mid page.

>The printer can never know how long it will take you to send an entire
>page, so it has to be willing to accept a region's worth at a time...
>with indeterminate pauses between them.

Ever see a laser printer that stop feeding paper so the computer can
catch up? I haven't.

>The same process can happen in the printer (e.g., send a PDL representation
>to the printer and let it create a portion of the page at a time). If
>the printer can *internally* render faster than the marking engine,
>then it only needs enough of a buffer to keep ahead of the marking engine.
>
>When printing text (using internal/downloaded font generation) and/or
>mixed text and limited graphics, this allows the printer to cheat on
>RAM requirements.
>
>My Phasers were 1200dpi and none of them had more than 200MB of RAM
>(and they can't pause and restart the marking engine once it's begun)

Which model Tek Phaser? The neighboring office had one. Long warmup
time, expensive wax bricks, wax in printed page would sometime melt or
crack, wax didn't stick to some types of paper, reds would fade, slow
printing, etc. Otherwise, I really liked it because the colors were
bright and accurate. I didn't know the Phasers could do 1200dpi. The
neighbor has one of the earliest Tek versions.

However, to be fair here, we were discussing the HP LJ 5MP, which is a
monochrome laser printer, not color. Color introduces a whole new set
of problems, including yet another increase in required RAM. Usually,
the printer manufacturers reduce capabilities in some way in order to
keep costs down. For example, the Brother mfc-9840cdw color laser
printer:
<https://support.brother.com/g/b/spec.aspx?c=us&lang=en&prod=mfc9840cdw_all>
prints at 2400 x 600 dpi because the base version only ships with
128MBytes of RAM. This can be increased to 640MB RAM, but the print
resolution remains the same. They're probably (not sure) using some
of the added RAM for rasterizing the entire page, and the rest for
print buffering, font smoothing, HPGL emulation(?), and Postscript 3
processing.

>So, whatever memory you have in your printer, by definition, is
>adequate for it to be able to print whatever it needs to print
>(subject to the caveats, below -- and above).

Well, that's a good general statement. Add to that how it's handled
if some of the processing is done in the computah (Truetype, JPEG
rendering), some of the processing in the printer (font smoothing, PCL
interpretation, Postscript, etc). At this time, there's quite a bit
of processing happening in both the computah and the printer. However,
the original problem was about trying to print todays complex
documents on a 1995 vintage laser printer. The computer might be
furiously computing, but the printer is still running at 1995 era
memory capacity, processing speeds, and capabilities. Without
sufficient printer RAM to build a bitmap of the printed page, a 24
year old printer is only going to print what was possible in 1995.

>> Inkjet printers only need to have a
>> strip the width of the print head stored in memory in order to print.
>
>But the printer only has to store what it *thinks* it needs to
>print. E.g., if you can't resolve 16 bits of grayscale, then you
>don't need to hold 16 bits of greyscale per pel.
>
>IIRC, color lasers print 4 different monochrome pages, in succession.

True. Each page has to be stored in printer RAM in order to print. To
save on RAM, my HP Color LaserJet MFP M477fnw will only do 600x600dpi,
or 38,400x600dpi in an enhanced dpi mode that I've never been able to
make work. RAM is 256 MB NAND Flash and 256 MB DRAM with no options
to add more RAM.

--
Jeff Liebermann jeffl@cruzio.com
PO Box 272 http://www.LearnByDestroying.com
Ben Lomond CA 95005-0272
Skype: JeffLiebermann AE6KS 831-336-2558

On Sun, 8 Aug 2021 11:46:28 -0700, Don Y <blockedofcourse@foo.invalid>
wrote:

>On 8/8/2021 11:28 AM, Don Y wrote:
>> On 8/8/2021 11:05 AM, Jeff Liebermann wrote:
>
>>> Trying to print an 8.5x11inch, 1200 dpi graphic, 16 bit gray scale,
>>> image is going to require quite a bit of computah RAM.
>>> 8.5x11 = 93.5 sq-in
>>> 1200x1200 pixels/sq-in = 1.44 Mpixels/sq-in
>>> 93.5 * 1.44M = 134,640,000 pixels per page
>>> Each pixel can have various shades of gray. The laser printers are
>>> either 8 or 16 bits/pixel (256 or 65,536) shades of gray.
>>> 134,640,000 pixels/page * 256 shades/pixel / 8
>>> = 4.3 MBytes/page
>>> memory required in the printer. Your 5MP printer can probably handle
>>> that. However 16 bits gray scale per pixel requires:
>>> 134,640,000 pixels/page * 65,536 shades/pixel / 8
>
>------------------------------------------------------^^^
>
>No, what you want is log2 of 65K -- 16.
>
>>> = 1.103*10^12 = 1.1 Terabytes of RAM.
>
>No. 134 Mpx and each needs 2 *bytes* to represent 65K shades of grey.
>So, ~270MB of RAM to represent a page.
>
>134MB at 256 shades of grey (@1200x1200 per color).

Oops. Y'er right. The 1.1 TBytes was a bit much and should have
failed even a basic sanity check, which I didn't bother doing. I used
to write my comments, save them, wait at least an hour, and review
what I wrote. That will usually catch most of my mistakes. I didn't
do that this time. Sorry.

>A 300dpi printer requires ~8MB for 256 shades of grey and 16MB for 65K.
>
>So, 1200x1200 is 16 times the density (per unit area) which would
>be 16 times the memory.

Thanks for catching yet another of my math screwups. Maybe I should
give up on math and switch to something that requires less math, like
politics?

--
Jeff Liebermann jeffl@cruzio.com
PO Box 272 http://www.LearnByDestroying.com
Ben Lomond CA 95005-0272
Skype: JeffLiebermann AE6KS 831-336-2558

On 8/8/2021 7:25 PM, Jeff Liebermann wrote:
> On Sun, 8 Aug 2021 11:28:43 -0700, Don Y <blockedofcourse@foo.invalid>
> wrote:
>
>> On 8/8/2021 11:05 AM, Jeff Liebermann wrote:
>
>> No, that's a naive expectation. You don't need to keep an entire
>> page in RAM to be able to generate a page *in* the computer.
>>
>> You treat the page as a series of smaller regions (the top 1", the
>> next 1", etc.) and define a clipping region into which you "draw"
>> the page.
>> Then, ship that portion off to the printer.
>> Then, move the clipping region to the next region and repeat the process.
>
> That works nicely for printers that can tolerate start/stop paper feed
> operation, such as inkjet printers and plotters. Laser printers need
> an uninterrupted paper feed (to keep the fuser from setting fire to
> the page when the paper stops) which is not likely to happen if the
> computah has to send parts of a page to the printer. Even with some
> buffering, any interruption in data transfer will cause a misprint
> simply because the paper cannot stop moving in mid page.

No. This addresses how the *driver* in the PC can build a page without
requiring a full page of bitmap memory. It computes *part* of a page
then ships that off to the printer. The printer sees that it doesn't
yet have the full page so it just waits.

The *printer* accumulates the page in these chunks ("regions").

When the driver has finished processing the last chunk/region
and sent it off to the printer, the printer notices that it now
has a complete page and starts the marking engine.

This is no different than if the driver had computed an entire
bitmap of the page... but in shipped it to the printer in "bursts".

>> The printer can never know how long it will take you to send an entire
>> page, so it has to be willing to accept a region's worth at a time...
>> with indeterminate pauses between them.
>
> Ever see a laser printer that stop feeding paper so the computer can
> catch up? I haven't.

Reread what I wrote.

Dear printer. Let's pretend I have 1200*11" rows of dots in this gigantic
buffer. Here is the first row. Now, I'm going to pause for a moment while
I check my horoscope. OK, here comes the next row. Hmmm, maybe I'll get
a bite to eat. Here's the next row, after that.

Etc. Until 1200*11 rows have been sent.

How is that any different, from the printer's perspective, than:

Dear printer. Let's pretend I have 1200*1" rows of dots in this small
buffer. Here is the first row. Now, I'm going to pause for a moment while
I check my horoscope. OK, here comes the next row. Hmmm, maybe I'll get
a bite to eat. Here's the next row, after that. For a total of 1200 rows.

OK, hold on a second (or, maybe a MINUTE!), I'm trying to sort out the
*next* 1200*1" chunk of stuff for you. Here is the first row of that next
group of rows...

Lather, rinse, repeat -- until 1200*(1+1+1+1+1+1+1+1+1+1+1)" rows of
data have been sent.

In the first case, the computer/driver had to buffer the entire page
of rasterized pels. In the second, the computer/driver only had to
deal with 1/11th of that.

>> The same process can happen in the printer (e.g., send a PDL representation
>> to the printer and let it create a portion of the page at a time). If
>> the printer can *internally* render faster than the marking engine,
>> then it only needs enough of a buffer to keep ahead of the marking engine.
>>
>> When printing text (using internal/downloaded font generation) and/or
>> mixed text and limited graphics, this allows the printer to cheat on
>> RAM requirements.
>>
>> My Phasers were 1200dpi and none of them had more than 200MB of RAM
>> (and they can't pause and restart the marking engine once it's begun)
>
> Which model Tek Phaser? The neighboring office had one. Long warmup
> time, expensive wax bricks, wax in printed page would sometime melt or
> crack, wax didn't stick to some types of paper, reds would fade, slow
> printing, etc. Otherwise, I really liked it because the colors were
> bright and accurate. I didn't know the Phasers could do 1200dpi. The
> neighbor has one of the earliest Tek versions.

I had several. I recall an 840, 860, 8200 (I think the 8200 had a duplexer)

The printers (and ink!) were all rescues so the cost that I bore was just
that of operation.

But, I'd typically only print a single publication (I prepare formal
documentation) and then power it down. So, lots of overhead, lots of
wasted ink in the powerup "purge" cycle. And the smell of burnt crayons.

But the results were excellent -- like right out of a magazine! Can't
get that same "finish" from a laser or inkjet.

> However, to be fair here, we were discussing the HP LJ 5MP, which is a
> monochrome laser printer, not color. Color introduces a whole new set
> of problems, including yet another increase in required RAM. Usually,
> the printer manufacturers reduce capabilities in some way in order to
> keep costs down. For example, the Brother mfc-9840cdw color laser
> printer:
> <https://support.brother.com/g/b/spec.aspx?c=us&lang=en&prod=mfc9840cdw_all>
> prints at 2400 x 600 dpi because the base version only ships with
> 128MBytes of RAM. This can be increased to 640MB RAM, but the print
> resolution remains the same. They're probably (not sure) using some
> of the added RAM for rasterizing the entire page, and the rest for
> print buffering, font smoothing, HPGL emulation(?), and Postscript 3
> processing.

If you have control over the driver, then you can split the load between the
driver and printer. E.g., render one "color" at a time in the PC and
ship it over.

My 8200 had a small (10G) laptop drive so I could install font libraries
in the printer -- instead of having to embed the font definitions in
the document being printed.

>> So, whatever memory you have in your printer, by definition, is
>> adequate for it to be able to print whatever it needs to print
>> (subject to the caveats, below -- and above).
>
> Well, that's a good general statement. Add to that how it's handled
> if some of the processing is done in the computah (Truetype, JPEG
> rendering), some of the processing in the printer (font smoothing, PCL
> interpretation, Postscript, etc). At this time, there's quite a bit
> of processing happening in both the computah and the printer. However,
> the original problem was about trying to print todays complex
> documents on a 1995 vintage laser printer. The computer might be
> furiously computing, but the printer is still running at 1995 era
> memory capacity, processing speeds, and capabilities. Without
> sufficient printer RAM to build a bitmap of the printed page, a 24
> year old printer is only going to print what was possible in 1995.

I'm running 24 year old printers. With 66MB of RAM, they can easily store
an image of a page (they even have a "page protect" feature that sets
aside a page's worth of RAM to cover worst case prints).

Building the page's image can be costly (time). But, no reason it can't
be done in the printer *or* the driver.

I suppose one could create a *representation* of a page that takes
more memory than the bitmap of the page. In which case, a smart
driver should know to prepare a bitmap instead of sending the
PDL (PCL, PS, etc.) to the printer.

>>> Inkjet printers only need to have a
>>> strip the width of the print head stored in memory in order to print.
>>
>> But the printer only has to store what it *thinks* it needs to
>> print. E.g., if you can't resolve 16 bits of grayscale, then you
>> don't need to hold 16 bits of greyscale per pel.
>>
>> IIRC, color lasers print 4 different monochrome pages, in succession.
>
> True. Each page has to be stored in printer RAM in order to print. To
> save on RAM, my HP Color LaserJet MFP M477fnw will only do 600x600dpi,
> or 38,400x600dpi in an enhanced dpi mode that I've never been able to
> make work. RAM is 256 MB NAND Flash and 256 MB DRAM with no options
> to add more RAM.

That's yet another reason to let someone else deal with making the
right printer available for use! Keep kit on hand for the "simple"
and common things. Let someone else worry about anything fancier!

On Sun, 8 Aug 2021 20:38:27 +0200, Gerhard Hoffmann <dk4xp@arcor.de>
wrote:

>Am 08.08.21 um 20:05 schrieb Jeff Liebermann:
>> On Sat, 07 Aug 2021 17:45:50 -0400, Joe Gwinn <joegwinn@comcast.net>
>> wrote:
>>
>>> On Sat, 07 Aug 2021 13:08:15 -0700, Jeff Liebermann <jeffl@cruzio.com>
>>> wrote:
>>>
>>>> On Fri, 06 Aug 2021 10:59:34 -0400, Joe Gwinn <joegwinn@comcast.net>
>>>> wrote:
>>>>> Are the gears made of nylon, or delrin?
>>>>> Delrin is claimed to be suitable for press fits, but not nylon.
>>>>
>>>> My guess(tm) would be glass reinforced nylon. The teeth are stronger
>>>> than acetal (delrin) and seems like the right color. However, since
>>>> the HP plastic shrank instead of enlarged when it absorbed moisture,
>>>> my guess(tm) is that it's a mix of nylon and something else intended
>>>> to compensate for expansion due to water absorption. I guess they
>>>> added too much.
>>>
>>> Nylon is most likely. There may also have been a plasticiser of some
>>> kind, which evaporated. Molded plastic gears are an optical illusion
>>> anyway.
>>
>> Nope. Plasticizer is what is used in various compounds, including
>> nylon, to make them soft and flexible. That's the last think I would
>> want to see in a toothed gear. Plasticizer can be used to reduce
>> friction in gears, but that's not needed in the HP8640B, where the
>> gears are rarely rotated and then only by hand.

>The HP 8640B now has the problem that its Delrin gears loose teeth.
>There is a ham in Italy who makes new wheels from brass.
>Gerhard

Are you sure the gears are delrin and not nylon? I'm not sure and
don't have any loose gears handy to check.
"HP 8640B Gears - Nylon or Delrin? DeVries says Nylon! "
<https://groups.io/g/HP-Agilent-Keysight-equipment/topic/69926273>

From India:
<https://www.ebay.com/itm/154054911606>

Cast resin gears:
<http://conradhoffman.com/HP8640B_gears.htm>

HP 8640b range & peak deviation gears replacement
<https://www.youtube.com/watch?v=MszkI32MFPE>

HP 8640 Service: Hints & Kinks
<https://www.ve7ca.net/TstH86.htm>
See item 2 and 3.

One of my HP 8640B generators on the left. A customers HP 8640B
partly disassembled on the right:
<http://www.learnbydestroying.com/jeffl/pics/HP8640B/hp8640b-fix.jpg>

I like my gear repair method better. Carefully save the pieces of
broken gear, which conveniently cracks in two pieces at the set screw
hole. Glue the pieces back together. Plug the set screw holes in the
gear and hub with wax. Use a large round file to enlarge the center
hole in the plastic until it fits over the brass insert. Glue the
parts together with epoxy and let harden. Clean out the set screw
threads with a tap and you're done.

--
Jeff Liebermann jeffl@cruzio.com
PO Box 272 http://www.LearnByDestroying.com
Ben Lomond CA 95005-0272
Skype: JeffLiebermann AE6KS 831-336-2558

Am 09.08.21 um 07:37 schrieb Jeff Liebermann:

>>> Nope. Plasticizer is what is used in various compounds, including
>>> nylon, to make them soft and flexible. That's the last think I would
>>> want to see in a toothed gear. Plasticizer can be used to reduce
>>> friction in gears, but that's not needed in the HP8640B, where the
>>> gears are rarely rotated and then only by hand.
>
>> The HP 8640B now has the problem that its Delrin gears loose teeth.
>> There is a ham in Italy who makes new wheels from brass.
>> Gerhard
>
> Are you sure the gears are delrin and not nylon? I'm not sure and
> don't have any loose gears handy to check.
> "HP 8640B Gears - Nylon or Delrin? DeVries says Nylon!"
> <https://groups.io/g/HP-Agilent-Keysight-equipment/topic/69926273>

I don't care. I just followed a discussion with a happy ending.
They said Delrin. But it kept me from buying one.

I mostly use R&S generators. I have a 8662A that is in its favorite
state: broken. Works from 0 to 160 MHz, does not from 160 to 320,
works again from 320 to 480, does not from 480.. and so on.
It will take some pressure to make me open it. Or boredom. Or someone
who says: "That's easy! mine had that, too..." :-)

Cheers, Gerhard

On Sat, 7 Aug 2021 15:32:39 -0700 (PDT), Lasse Langwadt Christensen
<langwadt@fonz.dk> wrote:

>søndag den 8. august 2021 kl. 00.23.59 UTC+2 skrev jla...@highlandsniptechnology.com:
>> On Sat, 7 Aug 2021 17:45:33 -0400, Phil Hobbs
>> <pcdhSpamM...@electrooptical.net> wrote:
>>
>> >Lasse Langwadt Christensen wrote:
>> >> fredag den 6. august 2021 kl. 11.51.34 UTC+2 skrev Don Y:
>> >>> On 8/6/2021 2:15 AM, Martin Brown wrote:
>> >>>> On 05/08/2021 16:22, Don Y wrote:
>> >>>
>> >>>>> Exactly. And, its relatively easy to convince a jury (likely
>> >>>>> comprised of equally "idiotic" people) that they could envision
>> >>>>> themselves doing something similar.
>> >>>>>
>> >>>>> "Why *can't* I set the ladder on cow shit? What *else* can't I do
>> >>>>> with it? Why didn't you tell me this ladder was so damn USELESS???"
>> >>>>
>> >>>> It is the sort of legalistic no win no fee game that forces people in the USA
>> >>>> to have "open other end" stamped on the base of glass bottles.
>> >>>>
>> >>>> In the UK we still have the concept that if someone does something terminally
>> >>>> stupid with equipment then that is their problem. Though increasingly we are
>> >>>> getting product warning to not do insane things!
>> >>> Who decides what is "terminally stupid"? Or, even marginally stupid?
>> >>>
>> >>
>> >> https://en.wikipedia.org/wiki/Bonus_pater_familias
>> >>
>> >>> Flour (for baking) comes with a warning to not eat it *raw*! (WTF? Who would
>> >>> find the taste of raw flour appealing??). Is there a risk that someone will
>> >>> die or suffer serious injury from doing so? (I'm *asking* as I don't know)
>> >>
>> >> flour potentially contain germs like E. coli, Salmonella, and Listeria
>> >>
>> >
>> >And mixing it with warm water (or maybe milk and sugar) and leaving it
>> >out for a bit can help it get going, if there's any left alive.
>> Some sourdough cultures have been fermenting for over a century. I
>> guess the good bugs kill off the bad ones.
>
>or the good bugs "poop" kills off the bad bugs, that's how the many kinds
>of food preserved with fermentation came about
>
>in the case of sourdough, once it is baked they are all dead

The giant Tartine loaves are heavy and sticky inside and really hard
to cut. The knife comes out gloppy. The insides don't seem to be baked
100%.

mandag den 9. august 2021 kl. 19.14.51 UTC+2 skrev John Larkin:
> On Sat, 7 Aug 2021 15:32:39 -0700 (PDT), Lasse Langwadt Christensen
> <lang...@fonz.dk> wrote:
> >sųndag den 8. august 2021 kl. 00.23.59 UTC+2 skrev jla...@highlandsniptechnology.com:
> >> On Sat, 7 Aug 2021 17:45:33 -0400, Phil Hobbs
> >> <pcdhSpamM...@electrooptical.net> wrote:
> >>
> >> >Lasse Langwadt Christensen wrote:
> >> >> fredag den 6. august 2021 kl. 11.51.34 UTC+2 skrev Don Y:
> >> >>> On 8/6/2021 2:15 AM, Martin Brown wrote:
> >> >>>> On 05/08/2021 16:22, Don Y wrote:
> >> >>>
> >> >>>>> Exactly. And, its relatively easy to convince a jury (likely
> >> >>>>> comprised of equally "idiotic" people) that they could envision
> >> >>>>> themselves doing something similar.
> >> >>>>>
> >> >>>>> "Why *can't* I set the ladder on cow shit? What *else* can't I do
> >> >>>>> with it? Why didn't you tell me this ladder was so damn USELESS???"
> >> >>>>
> >> >>>> It is the sort of legalistic no win no fee game that forces people in the USA
> >> >>>> to have "open other end" stamped on the base of glass bottles.
> >> >>>>
> >> >>>> In the UK we still have the concept that if someone does something terminally
> >> >>>> stupid with equipment then that is their problem. Though increasingly we are
> >> >>>> getting product warning to not do insane things!
> >> >>> Who decides what is "terminally stupid"? Or, even marginally stupid?
> >> >>>
> >> >>
> >> >> https://en.wikipedia.org/wiki/Bonus_pater_familias
> >> >>
> >> >>> Flour (for baking) comes with a warning to not eat it *raw*! (WTF? Who would
> >> >>> find the taste of raw flour appealing??). Is there a risk that someone will
> >> >>> die or suffer serious injury from doing so? (I'm *asking* as I don't know)
> >> >>
> >> >> flour potentially contain germs like E. coli, Salmonella, and Listeria
> >> >>
> >> >
> >> >And mixing it with warm water (or maybe milk and sugar) and leaving it
> >> >out for a bit can help it get going, if there's any left alive.
> >> Some sourdough cultures have been fermenting for over a century. I
> >> guess the good bugs kill off the bad ones.
> >
> >or the good bugs "poop" kills off the bad bugs, that's how the many kinds
> >of food preserved with fermentation came about
> >
> >in the case of sourdough, once it is baked they are all dead
> The giant Tartine loaves are heavy and sticky inside and really hard
> to cut. The knife comes out gloppy. The insides don't seem to be baked
> 100%.

as long as it reaches something like 70'C for a few minutes