Musings and comments about our common interest
If you have bought a HP41C, chances are that you have found a machine that has a broken back half. Typically, this happens when the top screw supports break under the tension required to bring together the two halves, putting presure on the foam under the flex circuit. A strong pressure is required for the calculator to work properly, as all of us that have assembled calculators remember.
The problem when repairing the broken supports is that it is very difficult to paste again the broken pieces - and when you do it, they tend to break easily. When you see a completely broken post, it is easy to see that the adherence area is very small. So we decided to find a solution. We designed a piece that can be adhered over a wider surface, that can be used for both sides or one side, and that can integrate woth the original battery-port assembly or with our new 3D-printed model.
This is how a partial repair looks like: first we cut the part we need:
Then, we need to remove all rests of the original screw support with a Dremel machine or similar. Then, we put the part in place. The way it is designed, it can only be put on way and it also serves to keep it in place. Then, we paste it and keep together with a clothes peg or similar overnight. And this how it looks with the repaired piece:
The part will fit under the battery connector module. Also, the way it is designed, it helps to keep a strong connection between the module and the circuit, even if for some reason the glue doesn't hold.
We have now 3 more circuits to be built on HP41CV bodies. These were just uploaded with the very latest versions of all modules.
The news is that we have a Time Module available to create the ultimate classic calculator.
The first order will be shipped with an original HP41C oversized vinyl case. Next units will go with leather-like new cases.
For those that are not aware of the capabilities of this machine, here are some of its features:
Please see all options in the HP41CL page. Prices start at 625 EUR.
We have just found a Time Module - which will be available for the HP41CLs that we have for sale!!
One of the common problems that plague old HP41c is the breaking of the upper screw supports in the back side of the calculator. The lack of pressure thus causes the battery circuit to not make full contact - and the calculator doesn't work. The repair is quite difficult, since the contact area for adhering the broken plastic is very limited. We have created a piece that fits below the battery contact assemble, has enough surface to adhere to the back side, and can even substitute for the broken stands that fix the battery contact assembly:
You can remove the rests of the two screw supports of the case and glue this piece in place. If one support remains in place, you can cut the piece in half and just glue one part of the piece. Then you can put the battery assembly (original or new) on top of it, guided by the elevated parts, since it it designed to fit as below:
We will now work on the process: find the right glue, how to remove the rest of the screw stands, etc., so it is not ready for ordering yet. Please email to firstname.lastname@example.org if you're interested on repairing your back-side with this piece. In any case, it is a good item for your repair box!
We know that many of you are put off by the difficulty of changing the battery connectors, due to the fact that you need to perform several steps:
For some time, and taking profit of the experience and steady hand of Ignacio Sánchez, we have offered preassembled kits: we took the plastic piece of a non-working calculator and performed the above steps - so that you did not have to. These parts sold immediately. The problem with it was that we were limited to the number of non-working calculators we could find - and that dried up very soon.
So we set up to create the plastic piece ourselves. We had to get familiar with the new CAD design programs - having used Autocad 25 years ago, there was a lot to be refreshed! Then we carefully measured the piece with a caliper, and once finished with the drawing, we sent it to a 3D printer company. Now we've finished the testing phase and are ready to deliver the first units.
This process has given us the chance to improve on the original (with softer bends and more extended adherence area), but also to create an alternate version (which can be seen on the left). If you have repaired several calculators, you surely have received one with the back side broken: the support for the screws in the back side has cracked. There is no repairing it with any kind of glue, as you know. Well, one of the versions of the battery assembly has a wide flat area, so that it can be glued to the back side and solve the problem with the broken calculator backs! The head of the screw will rest on the flat side of the piece, and the back side of the calculator is held in place due to the force of the glue over a wide area (not just slim borders as would be the case with the original piece)
Each unit will be sold for 62 € - with VAT and shipment included within European Union.
This is only to signal our customers that we have again some units or flexible circuit for repairing the battery bay of the HP41c family of calculators.
One of the most frequent requirements we're getting is to provide assembled repair kits for HP41C. The problem is that we depend on getting used, destroyed HP41c calculators to peel off the old flexible circuit, clean it and file off the rivet remains, and adhere the new flexible circuit to the plastic base. The issue is the first part - getting the part to perform the work.
Once you have used the part, your calculator cannot work - so you can only get it from damaged calculators that you're going to use as part donors. This is the reason that we don't get more of these assembled repair kits.
On the other hand, it really saves a lot of time and patience to avoid folding and collating the sticky flexible circuit - I always wonder whether it is going to work properly once finished. The assembled part takes the doubt off the process!
I will post here a message once I get more assembled repair kit parts.
I spent some quality time during this weekend to get the oscilloscope working. As I said in previous installments of the blog, the most likely culprit of the problem was the VRAM chip. These were built with a battery inside, and were used to keep the calibration variables. If it is not working, it not work properly until calibrated - but the calibration would then last only until powered off - and you'd have to recalibrate each time! Mine was even worse - it wouldn't show a trace. Running the test routines it failed on the VRAM (as expected) but also in the AD and DA tests.
I disassembled the machine, which was extrmely clean inside. The disassembly procedure was well explained in the service manual, and I managed to get the beautiful main board, equipped with a complete computer of the time - starting with the Morotola 68000 microprocessor (way more powerful than the 80X88 that equipped the primitive IBM PCs. The 68000 processor was also used in the Apple Macintosh)
The problem was that the integrated circuits were directly soldered to the board - no sockets in sight. Then it was risky to apply too much heat to solder or unsolder the chips.
After all the effort, we switched on the machine. It worked. Then we proceed to the recalibration as per the user manual (the machine is self-calibrating, which saves a lot of time getting the proper equipment to calibrate it. You only need a cinch-to-cinch connector to run from one of the calibrating outputs in the back of the oscilloscope to one of the four channels)
After the recalibration, we powered off and on again, and run the self-tests. This is what we got:
This meant that we had succeeded. This unit is a little bit battered on the outside, but excellent inside - I hadn't seen so clean a machine since a long time ago. Built like a tank, too.
I then proceeded to see the time response of the filters of the new Meridian Explorer 2 DAC, with MQA. I could then see the ringing of the filters, which is supposed to be all after the pulse (as opposed to Linear phase filters, which "ring" before and after the event.). These filters are supposed to be nicer on the ear. Lets see them applied to a sqare wave:
I already commented in a previous blog that I was repairing an 54501a oscilloscope. I was impressed by the quality of its circuit and general build; but also its logic was miles ahead of what I was used with analog oscilloscopes. Moreover, you could print out a screen through HP-IB to a suitable printer (that we also have).
The 100 MHz bandwidth may sound limited nowadays, with a lot of equipment running in the gigahertz range, from calculators to computers; on the other hand, it is plenty enough for any digital or analog audio applications, and this is what I am planning to use it for. Nowadays, with the advent of the MQA audio codecs, it will be priceless to perform all kind of analysis on how different waveforms are encoded.
As I have let the unit in the mountains' laboratory, I am not able to post any pictures - but i will during this weekend. The problem this unit seems to have is that it has lost the capacity to store a calibration. It used a vram memory chip that included its own battery, and it has run dry after all these years. The solution is to unsolder the old chip and solder in a replacement. This is a 5 euro chip, so it is a quite safe investment. Apparently, once this problem is sorted, it can be self-calibrated and the new values stored. Then -fingers crossed - all self tests will pass again.
I'll keep you posted. Please also email me if you have a similar machine to repair, or are looking for similar equipment.
I was always surprised to see the wide price differences between HP vintage oscilloscopes and HP spectrum analyzers from the same period. So it was with excitement that I got a HP 8591E spectrum analyzer for our shop.
Nowadays most of the current oscilloscopes have some kind of spectrum analysis tools - always performed as FFT (fast fourier transforms) of the waveform readings. These machines use a completely different method. I will copy from wikipedia so as to be able to explain it better:
A swept-tuned analyzer uses a superheterodyne receiver to down-convert a portion of the input signal spectrum to the center frequency of a narrow band-pass filter, whose instantaneous output power is recorded or displayed as a function of time. By sweeping the receiver's center-frequency (using a voltage-controlled oscillator) through a range of frequencies, the output is also a function of frequency. But while the sweep centers on any particular frequency, it may be missing short-duration events at other frequencies.
In future installments I will discuss this wonderful machine - complementary but completely different to your oscilloscope.