Looking for HP41CX Advice

[Neve]

Hi,

I have a HP41CX lying around that doesn’t see much use since I turned my older “Tall-Key” 41C into a CL a few years back.

Would you turn it into a CL as well (provided boards are available)? Or are there any other options you would consider, besides selling it.

I do like the LCD on the CX more than the one on the “Tall-Key”. It’s brighter with more contrast. It looks much better, particularly in the evening when there is less light.

PS: Also, I noticed the the stock CX turns on a few hundred milliseconds faster than the HP41CL). I don’t know why. If anyone knows, I’d be interested in knowing the reason. Is it due to the different LCD?

Thanks

[rprosperi]

(08-04-2021 04:55 PM)Neve Wrote:  PS: Also, I noticed the the stock CX turns on a few hundred milliseconds faster than the HP41CL). I don’t know why. If anyone knows, I’d be interested in knowing the reason. Is it due to the different LCD?

How are you timing that, do you have a calibrated eye?

If the turn-on time is the same regardless of CL TURBO speed, I'd say it probably is the LCD, but if not, and it does vary with TURBO speed (presumably quicker with speed increase) then it's likely because the CL has a lot more to do than a plain 'ol 41CX, checking CL configuration, etc. But Monte will see this and offer better insights.

[Neve]

(08-04-2021 07:36 PM)rprosperi Wrote:  

(08-04-2021 04:55 PM)Neve Wrote:  PS: Also, I noticed the the stock CX turns on a few hundred milliseconds faster than the HP41CL). I don’t know why. If anyone knows, I’d be interested in knowing the reason. Is it due to the different LCD?

How are you timing that, do you have a calibrated eye?

Yes, definitely.

(08-04-2021 07:36 PM)rprosperi Wrote:  If the turn-on time is the same regardless of CL TURBO speed, I'd say it probably is the LCD, but if not, and it does vary with TURBO speed (presumably quicker with speed increase) then it's likely because the CL has a lot more to do than a plain 'ol 41CX, checking CL configuration, etc. But Monte will see this and offer better insights.

Whatever the TURBO settings, it’s the same. Even with the MMU cleared and disabled, TURBO at 0, no difference.
The turn on is just slower. Anybody can easily see the difference. It’s probably about 300-400ms slower, I’d say.
It’s not a problem at all, I was just wondering why that could be.
It could be a slower responding LCD, or just that the CL has more stuff to check. EXT Memory maybe?? 600 instead of the stock 124 registers on the CX??

[Sylvain Cote]

41CL MMU configuration can changes startup and/or operation speed a lot.
The best way to validate startup speed is to deactivate the 41CL MMU and then do the comparison.
If you have similar speed, then you know that the slowness is coming from the ROM's that you have mapped on the 41CL.
Sylvain

[Sylvain Cote]

(08-04-2021 07:50 PM)Neve Wrote:  Whatever the TURBO settings, it’s the same. Even with the MMU cleared and disabled, TURBO at 0, no difference.
The turn on is just slower. Anybody can easily see the difference. It’s probably about 300-400ms slower, I’d say.
It’s not a problem at all, I was just wondering why that could be.
It could be a slower responding LCD, or just that the CL has more stuff to check. EXT Memory maybe?? 600 instead of the stock 124 registers on the CX??

I just did a power on startup test with a 41CX-HN (with 600 XM in it) against a 41CLv5 (MMU disabled) and effectively the 41CX is a little faster at bringing up the 0.0000.
But this is cosmically counter balanced with the 50x turbo of the 41CL.
Like Robert said, Monte is the only who can answer this question.

[Neve]

(08-04-2021 07:51 PM)Sylvain Cote Wrote:  41CL MMU configuration can changes startup and/or operation speed a lot.
The best way to validate startup speed is to deactivate the 41CL MMU and then do the comparison.
If you have similar speed, then you know that the slowness is coming from the ROM's that you have mapped on the 41CL.
Sylvain

As I wrote in my previous post, even with a MEMORY LOST, MMU CLEARED and disabled, TURBO 0, it’s still slower.
It’s hard to describe, but it looks like the display is popping up in a spongy way, or rather “fading in”, if that makes any sense; whereas the CX is almost instantaneously showing up.

I could almost describe it as a capacitor slowly charging up, whereas on the CX the charge seems instantaneous.

[Neve]

(08-04-2021 08:00 PM)Sylvain Cote Wrote:  

(08-04-2021 07:50 PM)Neve Wrote:  Whatever the TURBO settings, it’s the same. Even with the MMU cleared and disabled, TURBO at 0, no difference.
The turn on is just slower. Anybody can easily see the difference. It’s probably about 300-400ms slower, I’d say.
It’s not a problem at all, I was just wondering why that could be.
It could be a slower responding LCD, or just that the CL has more stuff to check. EXT Memory maybe?? 600 instead of the stock 124 registers on the CX??

I just did a power on startup test with a 41CX-HN (with 600 XM in it) against a 41CLv5 (MMU disabled) and effectively the 41CX is a little faster at bringing up the 0.0000.
But this is cosmically counter balanced with the 50x turbo of the 41CL.
Like Robert said, Monte is the only who can answer this question.

OK. My setup is a bit different, I guess…

[Neve]

The LCD was really a simple observation, not a real issue or anything.
Any piece of advice regarding my original “existential life threatening” question on what to do with my CX? LOL

[Sylvain Cote]

(08-04-2021 08:09 PM)Neve Wrote:  The LCD was really a simple observation, not a real issue or anything.

No problem, I never said it was either.

(08-04-2021 08:09 PM)Neve Wrote:  Any piece of advice regarding my original “existential life threatening” question on what to do with my CX? LOL

Assuming that your CX is a fullnut, I would keep it as a backup unit of the 41CL main unit.

[aurelio]

(08-04-2021 08:09 PM)Neve Wrote:  Any piece of advice regarding my original “existential life threatening” question on what to do with my CX? LOL

Hi Neve, please send me the CX (a present) for next birthday, in january:
you' ll have a problem less to manage and I'll increase my 41's collection for free

Forgive my irony please..
Did you never consider to riconvert the CX in a CL (you will have a CL with a better display brightness), and if are not orryfied by Frankestein, a CX “Tall-Key”...
anyhow you can always restore if you need the original “Tall-Key” 41C
or simply restore direclty your “Tall-Key” 41C , leaving the CX board as spare or selling it on TAS

[Neve]

(08-04-2021 09:09 PM)aurelio Wrote:  Hi Neve, please send me the CX (a present) for next birthday, in january:
you' ll have a problem less to manage and I'll increase my 41's collection for free

Send me your mailing address and your actual birth date.

(08-04-2021 09:09 PM)aurelio Wrote:  Did you never consider to riconvert the CX in a CL (you will have a CL with a better display brightness), and if are not orryfied by Frankestein, a CX “Tall-Key”...
anyhow you can always restore if you need the original “Tall-Key” 41C
or simply restore direclty your “Tall-Key” 41C , leaving the CX board as spare or selling it on TAS

Yes I have considered it. But, for the sake of the posts, if I can avoid having to open it again, I’d rather leave it as it is. Posts were already damaged when I bought it many, many years ago, and it was a pain to fix. That was way before 3D printed kits were available to fix them.

Sylvain, PM sent.

[Sylvain Cote]

(08-04-2021 09:38 PM)Neve Wrote:  Sylvain, PM sent.

Neve, PM replied!

[Neve]

(08-04-2021 10:02 PM)Sylvain Cote Wrote:  

(08-04-2021 09:38 PM)Neve Wrote:  Sylvain, PM sent.

Neve, PM replied!

Thank you!
2 answers in PM.

[Monte Dalrymple]

(08-04-2021 08:00 PM)Neve Wrote:  

(08-04-2021 07:51 PM)Sylvain Cote Wrote:  41CL MMU configuration can changes startup and/or operation speed a lot.
The best way to validate startup speed is to deactivate the 41CL MMU and then do the comparison.
If you have similar speed, then you know that the slowness is coming from the ROM's that you have mapped on the 41CL.
Sylvain

As I wrote in my previous post, even with a MEMORY LOST, MMU CLEARED and disabled, TURBO 0, it’s still slower.
It’s hard to describe, but it looks like the display is popping up in a spongy way, or rather “fading in”, if that makes any sense; whereas the CX is almost instantaneously showing up.

I could almost describe it as a capacitor slowly charging up, whereas on the CX the charge seems instantaneous.

The YFNZ power-on polling point code is about 75 instructions, or roughly 12mS. The LCD bias voltages in the 41CL are generated using low-power op-amps (for source/sink operation) and I optimized for power consumption rather than speed. I suppose that it is possible that these bias voltages (1.1V and 2.2V) might be slower to reach their final values than in the original HP-41, but it's not something that I cared about or ever investigated.

[Neve]

(08-05-2021 12:56 AM)Monte Dalrymple Wrote:  

(08-04-2021 08:00 PM)Neve Wrote:  As I wrote in my previous post, even with a MEMORY LOST, MMU CLEARED and disabled, TURBO 0, it’s still slower.
It’s hard to describe, but it looks like the display is popping up in a spongy way, or rather “fading in”, if that makes any sense; whereas the CX is almost instantaneously showing up.

I could almost describe it as a capacitor slowly charging up, whereas on the CX the charge seems instantaneous.

The YFNZ power-on polling point code is about 75 instructions, or roughly 12mS. The LCD bias voltages in the 41CL are generated using low-power op-amps (for source/sink operation) and I optimized for power consumption rather than speed. I suppose that it is possible that these bias voltages (1.1V and 2.2V) might be slower to reach their final values than in the original HP-41, but it's not something that I cared about or ever investigated.

Thanks Monte. Indeed this is neither important nor impairing anything. That was just something I noticed.

[GreyUser]

The HP41 display drivers, lcd’s and their interconnects went through many design iterations over the years of the product lifecycle.

The original display driver chips (only on the early tall key model) were encapsulated with a black, molded square cap epoxied into place. These modules require a different timing cap on the motherboard. The HP circuit identifier is C2 and the original tall key driver needs a 0.01 uf cap. Later versions are 470 pf (as used on on the CL logic board). This change is documented in the combined 41C/CV/CX service manual.

In a normal 41, the indicator this cap is the wrong value for the driver is that there will only be visible characters when you hold down any key. Release the key and lcd will go blank. I don’t know what the behavior would be when a 41CL board is installed.

Later boards were encapsulated with a black epoxy blob on p/n 00041-80003. This was followed by part number 00041-60146 which were encapsulated with a grey epoxy blob. Note that these two boards are of different physical sizes with the later -60146 part being wider. This can cause problems when attempting to put an older driver board together with a newer lcd.

The original tall key lcd’s had a greenish tint to the polarizer with not so great contrast. The later lcd’s have a neutral, clear polarizer which IMO provides much better contrast. I don’t know if there are any speed differences in the drivers, given the hardware differences, I guess this could be confirmed by looking at the lcd drive signals. I would agree that the older lcd’s do appear to have a slower visual response, it may be the lcd, the driver or a combination of both.

If I had a choice of which to turn into a CL, a tall key unit or a CX, it would be hands down the CX. With the tall key variant, you’ve got the lcd issues, the typical poor tactile feedback of the 1 and 3 keys, a case bottom that was designed for motherboard attachment with nuts and perhaps the worst feature, motherboard interconnects via a silicone elastomer that cannot handle the supply current. Hopefully that was fixed by HP service at some point with the later C/CV gold interconnect.

The original tall key 41C’s were quite the homing pigeons and most made the trip back Corvallis more than once. It is exceptionally rare to find one in the wild with all original parts.

[Neve]

GreyUser,

Thanks for the explanation. I knew some of this stuff, but not as in-depth as you.

My Tall-Key has already been converted to a CL (V5 board) years ago. I think it was 2017. I haven’t opened it since, nor do I plan to unless I really need to work on it.
I do remember, however, that it doesn’t have the black square epoxied drivers. But I couldn’t tell you how they were; I just don’t remember. I do know that according to the CL installation instructions, I didn’t need to add a cap, otherwise I would have. But I didn’t.

I do know that it does have bolts to hold the CPU board down, and 2 gold plated spring loaded balls are visible in the power outlet slot. The 1 and 3 keys are just fine, I the overall feel of the keyboard is, to me, a bit nicer and a touch « klickier » than my CX.
The feel of the taller keys is more positive to the touch. The drawback is that the alphabet is less visible because of the steeper angle. Not really an issue since the position of the keys became second nature, but still…
The original motherboard sports only 2 transistors instead of the 3, I think, that were present on earlier models. The unit’s serial number starts with 2027A…. MADE IN U.S.A. But it’s definitely not a very early one.
The LCD itself is ok. It’s a bit darker than the CX, for sure.
I don’t think there’s anything I can do. But admittedly, the LCD on the CX does look nicer.

[Massimo Gnerucci]

(08-08-2021 07:05 PM)GreyUser Wrote:  The original tall key 41C’s were quite the homing pigeons and most made the trip back Corvallis more than once. It is exceptionally rare to find one in the wild with all original parts.

All of my TK 41s are working, with original parts.
Call me lucky. :)

[Neve]

(08-08-2021 09:24 PM)Massimo Gnerucci Wrote:  

(08-08-2021 07:05 PM)GreyUser Wrote:  The original tall key 41C’s were quite the homing pigeons and most made the trip back Corvallis more than once. It is exceptionally rare to find one in the wild with all original parts.

All of my TK 41s are working, with original parts.
Call me lucky.

Mine was working perfectly too, and with all original parts as well, before I changed the motherboard for a CL one.

[GreyUser]

If your 1 and 3 keys has the same travel and click as the other keys, it is a probably a later production model where they must have changed the tolerances. It’s important to never over-tighten the nuts that secure the cpu board, it will reduce the 1-3 key travel and you’ll notice it. Ultimately, the fix is to abandon the nuts altogether and use the later design spacers over the posts to apply pressure on the cpu board from the case back.

If you choose to use spacers, which IMO is the best overall solution, you can fab them yourself from styrene stock available in hobby stores. When doing this, you must insure the lower case screws are 3/8” long, not the original 1/4”. The shorter screws will absolutely strip out and they also lead to low post fractures. On any 41C that I repaired that had nuts, I never put them back on, I installed spacers. HP had different size nuts for posts that had stripped but I always thought it was just bad solution chasing a bad design.

As for 1-3 tall key issue, I’ve never been able figure out the exact cause, other than for over-tightening the nuts. I’ve found that some machines just will never have the proper travel or feel, no matter what you do and others seem to be fine. The only correlation I’ve found is that the older the s/n, the more likely it will have the two keys without the usual travel and click feel.

Of all the HP keyboards, my favorite would be the 41. Best combination of travel and tactile feedback. After almost fifty years of HP use, I haven’t decided if I prefer the basic function keys on the left side or the right.