RWAP Software

Hello.
Now I have time, I decided to test the interface "divide +", which was bought second hand through ebay (sn: Z-19). I have 3 Spectrums and after checking that do not work as expected, I want to exhibit different behaviors.
First in a rubber Spectrum 48k 4B issue, "the divide plus" goes into an endless loop, which calls press space to detect storage devices, and does not let you do anything else.
Second, in another Spectrum 48k, 4B also issue "the divide plus" apparently correctly detects flash memory (32M), but pressing fatware NMI and enter the list of directories and files appear corrupt, and nothing works.

and Third, testing with a Spectrum +3, (closing the internal bridge correspodiente "divide plus"), flash memory correctly detected, the fatware works, and seems resiDos also works well.

So the question is: what happens to my splits plus? need an external power supply, because it seems that with the Spectrum +3, works OK, not the case in the other two cases.

I do not know the age of the firmware I have, it would be appropriate to change the GAL for the new date?

I need help, someone can tell me. Thank you.

I want to apologize for using the google translator, but I have little knowledge of English.

It could well be that you would benefit from the new GALs.

The problem with corruption on the file browser when you press the NMI button is normally the compact flash card is not 100% compatible - although you say it works without issue on the +3. I wonder if the compact flash card has been formatted as IDEDOS or is corrupt in some way - it may be worth reformatting the card as FAT-12 and try again.

It could also be your power supply - have you tried a different power supply with the rubber key Spectrums?

The Spectrum where the DivIDE Plus gets stuck in a loop suggests it may be the Z80 needs to be replaced (it may lack the important M1 line used by interfaces such as DivIDE Plus, PlusD and the Multiface).

Hello. When you say that no M1 line, you mean that the microprocessor does it internally? PCBs that "issue4" if coming from the socket properly z80 to the expansion connector. I have verified that both microprocessors are genuine Zilog Z80 (the trouble is that no socket).
I've done other checks, as I've gotten several spectrum and inves_spectrum +2 + (spanish version clone).
In inves_spectrum + the "divide +" works correctly with the power supply of the above. (I have only the power supplies of the speccy rubber), not whether to have fewer chips in pcb, this computer model has lower power consumption and therefore operates divides +.
Neither compact flash if they have a great diversity of electricity consumption, or are negligible.
In the Spectrum +2 A, (4 computers with a single power supply), the divide + works correctly.

A few days ago I have bought the firmware selector to divide +, and I
installed. I loaded the residos from a +3, and I spent one of the spectrum 48 with problems, the fatware continues with the same problem detection loop of CF, but if restarting and pressing "1" to access firmware selector, I residos change to this seems to work for some time (several minutes), until it begins to show corrupted data. And from here I have to change the split + to another computer.

Conclusions.
- "Divide +" appears to be well
- "CF 32M, appears to work well in +2 A, +3 and Inves +
questions:
-Remains to be tested with a good source to see if it is the power consumption problem.
-See if feasible desolder the Z80 to change, although Zilog, management will have the M1.
-It could be some kind of problem in the expansion bus, or on the PCB?.

I accept suggestions for finding the problem.

Hmm if both processors are Zilog ones, then they should have the M1 line (although it is not guaranteed) - yes, the M1 line is supplied by the chip itself.

At least we know the issue is not with the DivIDE Plus unit. It may be worth trying the upgraded GALs as these fix a couple of bugs in the original GALs, as well as offering enhancements.

The other option would be to check the coil is operating at the correct frequency. I have recently had to have two 48K ZX Spectrums recapped (new capacitors) as the coil was operating at too low a frequency. It seems the capacitors are drying out after all these years! That could also explain why the interface can appear to work initially and then stop working

I tried to measure the frequency between C46 + terminal and ground with an oscilloscope (seeking 114Khz expected) and it has been impossible because the signal is not stable; fluctuates a lot, it looks more like a PWM signal to a fixed . How should this sign? If not well I should replace the electrolytic capacitors and also the 2 transistors?. Measuring with a multimeter, I measured 0.4 V AC and 10V DC. Which should be the correct values​​?
Greetings thanks.

hello. As I mentioned in the previous post, the signal fecuencia DC DC converter is correct, 114KHz, I do not know that this signal can be important on the "divide +" while tensions have needs. On the other hand I have found that the Zilog Z80, no signal M1, and that is why the "divide" does not work. But this is history, because I changed the Z80 one with M1 and working properly.

The major problem to come later, the divide + is now dead and the other world has led to two Spectrums. The problem is that the connector of expansion of the divided + lost leading separator between positions 5 and 6 of the bus, and when I noticed was late. How is it possible that this separator is lost? Why was not glued or screwed?. The truth is that by not getting spectrum centered at times have been lucky and not passed anything, but now dead two Spectrums. I repaired one, changing one z80, 2 transistors, 1 ROM are dead, another faulty and defective ULA.

And now, I do not know if it can be repaired so economical "divide +". First put a spacer on the bus, and then depues I could not be broken. One symptom is that the computer will not start and the temperature of the GAL is single (A_GAL) up in a few seconds to burn your finger. Does that can fix just changing the 3 GALS, and with the new updated version, or do some chip SMD could be spoiled?

I hope news. Thank you.

Oops - I have checked with Ian Gledhill - the Zilogs are (in his opinion) the worst for missing the M1 line

As for the DivIDE+ I am surprised that the key has come out, as they are normally secured with glue (possibly it was overlooked on this one) - you may be lucky enough in just replacing the GAL chips - it is normally these which are affected first.

Hello.
I changed the 3 GALs that you sent me.
The "divide +" now at least do something more than before. But it does not work. There must be some other bad component. Not sure if it 29F040 firmware, RAM, a diode or other SMD component. I am able to change the SMD components, but do not want to change all. I need help locating the problem if you tell me where to measure and what values ​​I find.

I tested with a 32M CF already worked before, and rarely detected me. I also tested with a hard drive with your power supply, and either. Both work OK on PC. I checked the other 3 chips are socket (2x74LS245 and 74ALS573), and well. If I can see the switch firmware, but only sometimes, sometimes not reset, the other is blocked with yellow LED on. The fatware not work, and sometimes goes corrupt content, but before damage the CF work and as I said in the CF and PC HD well. The resiDOS not work. If I select ZX BASIC, restarts and crashes, displaying only symbol "©" of the message "© 1982 Sinclair .... / .. © 1982 Amstrad." And then can go into an infinite loop in the menu fatware of detection of IDE devices.

I have also seen that the "divide +" is interfering in any way on the quality of the image, being more prominent problem in the Spectrum 48K (imagine that having worst power supply), and less in the +2 A, is observed a dark horizontal, which slowly descends onto the screen, including the edge.
To Touch I have not seen any overly hot component.
Nor do I know if my PCB version is very old and perhaps updating GALs, does not work, I read diwIDE, and DIWI-225 in the PCB.

I hope news. Thank you.

After a few months without touching the interface divided + and after changing the GAL that you sold me, I've been testing, and I have not come to any clear conclusion. Either the EEPROM is corrupted or some other chip apart from the GAL is wrong. What I have seen is that no EEPROM resident software works (fatware, DEMFIR etc). However when I load the file "Residos.tap" in RAM, this system works more or less correctly, unless some program will crash and corrupt the RAM and have to reload as mentioned in the manual.

What do you think that is the fault? EEPROM or another chip? How I can load the software on the EEPROM? Original fatware programs I've tried not loaded. Do you suministraias the software to do it, or do you send me an EEPROM programmed, and I know that I would play desoldering the old and put the new.?

I expect to hear, my previous post was not answered.

I am sorry that I have missed your previous post.

It is peculiar that the DivIDE Plus presumably still works fine with the Spectrum +3, just not the plus models.

I presume that you remove the +3 jumper when in use on the original Spectrums, and that there is a battery in the unit.....

I can send you a replacement EEPROM if that might help - these are actually the square chip in a socket, so quite easy to remove with a small screwdriver in a corner. Can you PM me your name and address, so that I can ensure I send it to the correct person.

The other option would of course be to return the DivIDE Plus to me for testing here - presuming it has a fault, I could swap it with another unit for £15 plus postage (as you did not purchase the original through me) and send it to Piotr to have a look at and repair if possible.