1858-0044 ?

[teenix]

Did anyone ever find out a cross reference for this part?

It is used for the thermal print head, display and motor drivers on calculators like the 19C and 91, 92 and 97.

There are lots of mentions of cross references to the Allegro TPQ3904 which I think is wrong as these are rated at 200mA, and not sufficient considering the 500mA heads.

The Allegro TPQ2222A might be a better replacement as it has the same pinout and rated up to 600mA. Some are still around from Digikey for example, but expensive. There is a cheaper surface mount part but would need a carrier PCB to convert to through hole mounting.

cheers

Tony

[AndiGer]

Victor Toth mentions exactly 2N2222
http://rskey.org/CMS/exhibit-hall/?view=article&id=130
Andi

[brouhaha]

(09-28-2023 12:46 AM)teenix Wrote:  There are lots of mentions of cross references to the Allegro TPQ3904 which I think is wrong as these are rated at 200mA, and not sufficient considering the 500mA heads.

It's possible that a TPQ3904 would be satisfactory due to the duty cycle, but I wouldn't count on it.

Why not just use discrete PN2222A transistors? That's the TO-92 equivalent of the 2N2222A (TO-18 metal can).

Quote:There is a cheaper surface mount part but would need a carrier PCB to convert to through hole mounting.

If you did build a carrier board, four MMBT2222A surface mount transistors would be much cheaper than any quad transistor package.

[teenix]

(09-28-2023 09:25 PM)brouhaha Wrote:  If you did build a carrier board, four MMBT2222A surface mount transistors would be much cheaper than any quad transistor package.

I did design a carrier board some time ago with 3904 types but was delayed because of suitable connector pins, but I may have found some. If so, and I get time, I'll change to 2222.

cheers

Tony

[Jeff_Birt]

What about a ULN2002A, ULN2003A? The pinout is completely different, and you have 7 instead of 4 drivers in one package. But depending on how the original is wired it might be a viable option.

[brouhaha]

(09-28-2023 09:25 PM)brouhaha Wrote:  If you did build a carrier board, four MMBT2222A surface mount transistors would be much cheaper than any quad transistor package.

The Central Semicondcutor MPQ2222A (16 DIP) is still available, for around $5.15 in quantity one.. The individual transistor dissipation limit is 650mW, and the package dissipation spec is 1.9W. Since the saturation voltage is around 1V at 500mA, each transistor driving 500mA will dissipate 500mW, but of course at lower duty cycle the power dissipation will be lower. Worst case in practice would probably be the top and bottom lines of printing "0.000000000". Given the printhead return time, I'm guessing that the overall duty cycle is probably less than 30%, and less than that for all four transistors in one package. The package used for the motor driver is wired such that only one of the transistors should be drawing the full motor current.

The SOT-23 MMBT2222A has a much lower power dissipation rating, around 225 mW for FR-4 PCB with 1oz copper and a larger-than-usual copper pad for the collector. I think it will still be within spec if you use 1oz copper and put as big of a pad as you can for the collector, as the package is designed so that the collector terminal is where most of the heat winds up. Ideally divide the component side (other than the DIP pins at the edges of the adapter) into four large copper pads for the collectors, and use vias to wire up the bases and emitters on the circuit side.

[teenix]

The carrier board being so narrow, the transistor pins would probably sink heat through the board connecting pins, about 2mm away, to the circuit board below. The collector pin could be soldered direct. Perhaps if also socketed, maybe this would add to the thermal mass. How efficient this would all be though, I'm not sure, so would need some testing.

PBHV8110DA, NPN, 1A, 0.4V VceSat, 1.25W in a SOT-23. Still need to get rid of the heat though, 25mm square pads suggested at that power.

Probably the worst case for the print heads/motor would be low battery where the duty cycle gets higher. I did scope the printer drivers some time ago, but I've lost the images :-(

The PIK IC must have the drive circuitry incorporated inside as it directly connects to the transistor bases. I think the Anode Driver has an external resistor common to each segment drive on the Vcc pin (5).

cheers

Tony

[brouhaha]

(09-29-2023 10:22 PM)teenix Wrote:  PBHV8110DA, NPN, 1A, 0.4V VceSat, 1.25W in a SOT-23. Still need to get rid of the heat though, 25mm square pads suggested at that power.

That Vce(sat) spec seems impressive, until one looks at Note1, whcih says that it's at pulse width <=300us, duty cycle<=2%. If that characteristic isn't similar at a much higher duty cycle, then it's not going to be better than an MMBT2222A. Given how much more this costs than the MMBT2222A, it might be worth prototyping with this part and measuring what saturation voltage you actually observe.

Also if you have less Vce(sat) in your transistors, and want to run the printhead at the same current, the delta in voltage drop for the transistor has to be made up somewhere else, so the heat dissipation will be transferred there. That's the reason for not switching to nFETs, which would lower the dissipation in the transistors even more than the PBHV might.

The 1.25W power dissipation is for 1 in^2 copper pad, so the limit on the DIP PCB is going to be much lower, probably not too different than the power dissipation limit of the MMBT2222A.

Quote:The PIK IC must have the drive circuitry incorporated inside as it directly connects to the transistor bases. I think the Anode Driver has an external resistor common to each segment drive on the Vcc pin (5).

It probably does have beefier output transistors. In PMOS, pulling high works well, but you can't generally get much current pulling low. That's convenient for driving the bases of NPN transistors. I don't know how many mA they can drive without the Vout going too low, but I'd guess it's probably not much more than 15mA, and they probably count on the transistor hFE being 40 or better at the 500mA collector current. (That's the 2222A hFE spec.)

[Hiwi]

What about MPS650/651 or BC639

[teenix]

Hi all,

I constructed a board based on 4 x PN2222, shown next to a CMOS 14 pin IC. I haven't tried it yet as I have to unsolder original parts to swap it over.

I'll have a go when I get time.

cheers

Tony

[jklsadf]

The 1858-0044 transistor array is often listed as a 3904 in cross references, but is actually a long-obsolete part from RCA, the CA3138G. See page 228-229 of https://www.rsp-italy.it/Electronics/Dat...201978.pdf
For modern parts, the 2222 would be a closer match.

   
   

[brouhaha]

(12-11-2023 06:53 AM)jklsadf Wrote:  CA3138G

Thanks for sharing that! It's interesting that the Vce(sat) spec of the CA3138G is a LOT better than a 2N2222[A] or TPQ2222[A].

Just out of curiosity, what is your information source that cross-referenced 1858-0044 to CA3138G?

[jklsadf]

Hi, for V_CE(sat), the requirement for 1858-0044 is listed as V_CE(sat)=0.3V max for I_C=400mA and I_B=8mA from 0 to 50 °C, and V_CE(sat)=0.45V max for I_C=600mA and I_B=12mA (no temperature listed, so only at 25 °C ambient). It's unclear if these are supposed to be pulsed ratings. The h_FE is specified as min 70 and max 400 for I_C=500mA and V_CE=1V. V_BE is specified as 1.2V max for I_C=500mA and V_CE=1V. Power dissipation is listed as 500mW absolute max for each transistor, and I_C as 600mA absolute max (but I'm assuming HP didn't run any of these anywhere near the absolute max, at least not continuously).

Apparently GE and Motorola used to be second sources for 1858-0044, but I have no idea what part numbers. The specs for a 2N2222 aren't as good, but it seems like people have substituted them successfully for some applications. It might be possible to substitute a logic-level nFET as well.

As far as the source, I can't share the source for cross referencing 1858-0044 to CA3138G, since it's not publicly available info. (I probably shouldn't actually be sharing this info, but it's for a long-obsolete part (albeit with specs that are difficult to find modern substitutes for), and hopefully it might be useful to people trying to repair old equipment.)