Hello fellow builders.I thought that it would be a nice idea to start a thread on current FSBUS (ver 2.0.8) bugs, so that we identify them, group them and help Dirk fix them. I assume Dirk is getting a lot of emails so maybe an updated list of bugs will help him (and us :-) ).1. When trying to flash a PIC with a big ID (i.e. 29) the PIC is flashed but not answers to that ID. I tried flashing a KEY card with ID 29, and the PIC was actually flashed with ID 5. This is reproducable and discovered by Eddie.2. When any kind of input occupies BIT 3 or BIT 7 (subIDs 3, 7 ) any trigger of either turns on BOTH those bits. This occurs in every row/bank and on either input bit. I.e. if I connect a simple push button to subID 3 and leave 7 unconnected, when I press thebutton and monitor the key press both subIDs 3 & 7 become active. Same happens when the button is connected to subID 7 and leave subID 3 unconnected. Bot subIDs 3 & 7 become active.3. If you set up a rotary with type Alps (via redec) to increase/decrease a value (whichever), even if the step if set to 1, the value changes by a factor of 2.I.e. if step is set to 1, the value changing by 2, if value set to 10, value changes by 20, etc. This does not occur when using the emulated up/dn switches from within FSBUS.4. If you set up altitude to change by a step of 50 (for example to overcome the above problem), the value is ignored by FS and as soon as you stop turning the rotary it reverts to the previously set value in FS. This does NOT occur when the step is set to 100.5. When a rotary is assigned to speed, there seems to be a conflict between FS and the input of FSBUS. Either increasing or decreasing IAS, although the value displayed on the FS panel shows the change normally, the relative FSBUS display goes crazy. Display changes rapidly beetween the previous IAS value in FS and the new changed value from FSBUS. I can't really describe it properly, you will have to try for yourself to see what I mean.6. Phase-shift Rotary encoders when interfaced to FSBUS via a REDEC decoder actually work as an up/down knitter switch. Redec translates the rotary code and closes one contact on one direction and another contact on the other direction. Each contact is one (1) Bit. Currently such a setup occupies four (4) bits of a KEY bank, instead of just the two (2) required. I think this is a left over from older versions.That's all I have found so far. Please feel free to add the bugs you found or comment as appropriate. I would be a good idea if we could keep this post visible on the first page for a few days so that we gather all the information.Fly safe,George DorkofikisAthens, Greece

Hi GeorgeGood idea, I have another one.Only one rotary(via redec) will be recognized within version 2.08.Jan.

>1. When trying to flash a PIC with a big ID (i.e. 29) the PIC>is flashed but not answers to that ID. I tried flashing a KEY>card with ID 29, and the PIC was actually flashed with ID 5. >This is reproducable and discovered by Eddie.This you can check easily: When doing "read pic" on the flash dialog, the first 2 characters are the module type and the next two are the ID. So if you have something like 0110 (dont remember what module type id's are, so ignore the "01" there) - the PIC CID would be 16 (0x10 in hexadecimal)>2. When any kind of input occupies BIT 3 or BIT 7 (subIDs 3, 7>) any trigger of either turns on BOTH those bits. This occurs>in every row/bank and on either input bit. I.e. if I connect>a simple push button to subID 3 and leave 7 unconnected, when>I press the>button and monitor the key press both subIDs 3 & 7 become>active. Same happens when the button is connected to subID 7>and leave subID 3 unconnected. Bot subIDs 3 & 7 become>active.This sounds very much like a short circuit on your key module, check the relevant pins and trace the circuit board around those pins, and check if you can follow the logic and maybe you find a small short circuit.Not a FSBUS bug.>3. If you set up a rotary with type Alps (via redec) to>increase/decrease a value (whichever), even if the step if set>to 1, the value changes by a factor of 2.>I.e. if step is set to 1, the value changing by 2, if value>set to 10, value changes by 20, etc. This does not occur when>using the emulated up/dn switches from within FSBUS.Isnt Redec for the knitter-type? ALPS is a bit different, try with the knitter mode. Also, if the test buttons on the fsbus toolbar work, but your reded doesnt, sounds like a problem in redec, or your encoder.Of course you can try div: 2>5. When a rotary is assigned to speed, there seems to be a>conflict between FS and the input of FSBUS. Either increasing>or decreasing IAS, although the value displayed on the FS>panel shows the change normally, the relative FSBUS display>goes crazy. Display changes rapidly beetween the previous IAS>value in FS and the new changed value from FSBUS. I can't>really describe it properly, you will have to try for yourself>to see what I mean.Are you sure this is the autopilot IAS and not the airspeed indicator value? :) That could cause a bit of a conflict if one tries to move the airspeed gauge :)Best,Tuomas

Hi Tuomas>>1. When trying to flash a PIC with a big ID (i.e. 29) the>PIC>>is flashed but not answers to that ID. I tried flashing a>KEY>>card with ID 29, and the PIC was actually flashed with ID 5.>>>This is reproducable and discovered by Eddie.>>This you can check easily: When doing "read pic" on the flash>dialog, the first 2 characters are the module type and the>next two are the ID. So if you have something like 0110 (dont>remember what module type id's are, so ignore the "01" there)>- the PIC CID would be 16 (0x10 in hexadecimal)Thanks for the info. Sure works faster that way. But the problem is why if I set up to flash id 29, it flashes 5?.>>2. When any kind of input occupies BIT 3 or BIT 7 (subIDs 3,>7>>) any trigger of either turns on BOTH those bits. This>occurs>>in every row/bank and on either input bit. I.e. if I>connect>>a simple push button to subID 3 and leave 7 unconnected,>when>>I press the>>button and monitor the key press both subIDs 3 & 7 become>>active. Same happens when the button is connected to subID>7>>and leave subID 3 unconnected. Bot subIDs 3 & 7 become>>active.>>This sounds very much like a short circuit on your key module,>check the relevant pins and trace the circuit board around>those pins, and check if you can follow the logic and maybe>you find a small short circuit.>Not a FSBUS bug.It's not. I've checked! And also I have changed the KEY board withanother one where the subIDs are the one after the other so subID 7 is at the end of the pin header (pins 15/16) and subID 3 is at pins 5/6.However I did measure the board. No shortcircuits exist.>>>3. If you set up a rotary with type Alps (via redec) to>>increase/decrease a value (whichever), even if the step if>set>>to 1, the value changes by a factor of 2.>>I.e. if step is set to 1, the value changing by 2, if value>>set to 10, value changes by 20, etc. This does not occur>when>>using the emulated up/dn switches from within FSBUS.>>Isnt Redec for the knitter-type? ALPS is a bit different, try>with the knitter mode. Also, if the test buttons on the fsbus>toolbar work, but your reded doesnt, sounds like a problem in>redec, or your encoder.No. The Knitter-type don't require a redec. The redec actually translates the usual rotary encoder to a 'knitter like' output!This was ok until ver2.0.2. It's not the REDEC because it was working fine on ver 2. The problem appeared when the rotaries were groupped together. When the Rotary2 type existed it worked great!>Of course you can try div: 2Where? How? Hmmm... Did I miss that option?>>5. When a rotary is assigned to speed, there seems to be a>>conflict between FS and the input of FSBUS. Either>increasing>>or decreasing IAS, although the value displayed on the FS>>panel shows the change normally, the relative FSBUS display>>goes crazy. Display changes rapidly beetween the previous>IAS>>value in FS and the new changed value from FSBUS. I can't>>really describe it properly, you will have to try for>yourself>>to see what I mean.>>Are you sure this is the autopilot IAS and not the airspeed>indicator value? :) That could cause a bit of a conflict if>one tries to move the airspeed gauge :)Affirmative, pretty sure. I also checked the FSUIPC offset and it's the correct one (07E2 is I recall correctly).Regards,George DorkofikisAthens, Greece

>Hi George>>Good idea, I have another one.>Only one rotary(via redec) will be recognized within version>2.08.>>Jan.Hi jan,Hmmm... This problem was corrected since ver 2.0.4 (I think).I don't have this problem.Are you sure you connect them properly? As I said each rotary occupies 4bits so you must connect the first one on subIDs 0 and 1 (pins 1-2 & 3-4) and the other one at subIDs 4 & 5 ( pins 15-16 & 13-14).Caution. The subIDs (and pins) are not one after the other. This is amistake I always did and I had to ammend my board to make it simple.In the key PCB provided by Dirk the subIDs are in the order of 0 1 2 3 7 6 5 4 and not 0 1 2 3 4 5 6 7.To make it simple :subID 0 : Pins 1 - 2subID 1 : Pins 3 - 4subID 2 : Pins 5 - 6subID 3 : Pins 7 - 8subID 4 : Pins 15-16subID 5 : Pins 13-14subID 6 : Pins 11-12SubID 7 : Pins 9 -10Hope this helps,George DorkofikisAthens, Greece

>Thanks for the info. Sure works faster that way. But the>problem is why if I set up to flash id 29, it flashes 5?.That is a good question :)>However I did measure the board. No shortcircuits exist.Hm. Then I am out of ideas. Maybe someone else can shed a light on this.>No. The Knitter-type don't require a redec. The redec>actually translates the usual rotary encoder to a 'knitter>like' output!>This was ok until ver2.0.2. It's not the REDEC because it was>working fine on ver 2. The problem appeared when the rotaries>were groupped together. When the Rotary2 type existed it>worked great!Oh yea, I meant the same. I mean with redec you need to use the "Knitter" mode, not ALPS AFAIK, just like you apparently do.Sounds like a change/bug in fsbus then. Lets see what Dirk says.>>Of course you can try div: 2>Where? How? Hmmm... Did I miss that option?I think the Div stuff happens in the ini files now, you could of course copy the line from the fssomething.ini to myfssomething.ini (I forget the exact filename now) and change the div to 2 for that, if that works as a temporary solution.Tuomas

>Oh yea, I meant the same. I mean with redec you need to use>the "Knitter" mode, not ALPS AFAIK, just like you apparently>do.>Sounds like a change/bug in fsbus then. Lets see what Dirk>says.Nope, the 'Knitter' mode doesn't work with the Redec! I had to use'Alps' mode whereas on ver 2.0.4 I had to use the 'Conrad' option!That's why I say it's an FSBUS bug. I'm confident Dirk will find it.>>>Of course you can try div: 2>>Where? How? Hmmm... Did I miss that option?>>I think the Div stuff happens in the ini files now, you could>of course copy the line from the fssomething.ini to>myfssomething.ini (I forget the exact filename now) and change>the div to 2 for that, if that works as a temporary solution.I'll have a close look and see what I can work out. Thanks.George DorkofikisAthens, Greece

Using the analog board, let's say, flshed as CID 9, if I use a pot on this board, and auto-sense it, it comes up as if it were CID 1 (which is not, checked reading the code back). If I manually set it to CID 9 and RID 0 (for example), it won't work.I was trying to assign potentiometers to the throttle axis to eliminate the noise problem of using a hacked joystick circuitry.Any ideas?Cheers,Rafa.

This is what is in the protokol.pdfFestlegung der CID

Thanks a lot Jan,It worked! But now I am in trouble. What happens if I need more than 4 CIDS for keycards? Assuming i need to use 6 FSKey boards, what do I do to connect the other two? FSBUS supports 31 CIDs, but can it only accept 4 key and 4 analog boards?Thank you.Rafa.

RafaYou have 64 entry's on a key PCB. You can connect 256 switches etc. on 4 pcb's.What do you want to connect???????????Maybe Dirk can give the answer.Jan

Well, Lotsa switches, lots of rotary encoders, lotsa displays, lotsa leds, for a big 737 cockpit complete with overhead panel.;-)

>Well, >>Lotsa switches, lots of rotary encoders, lotsa displays, lotsa>leds, for a big 737 cockpit complete with overhead panel.>>;-)>>It's FS-*B*U*S* A BUS. Think "IDE-BUS" and "IDE-cable".

COM-1============[]=====[]=====[]=====[]				 KEY	KEY	KEY	LEDCOM-2============[]=====[]=====[]=====[]			   ANALOG  SERVO   LED	KEYCOM-3COM-4

Get the idea? Make flat cables that have several connectors. Those 4 connectors on the COM-card are interconnected anyway when you look at the layout.(yes, this has been discussed in the past, use the search ;))You can do one long cable with 31 connectors and hook all modules to the same connector in the COM-card if you want - that's why the PICs have the unique ID.Tuomas

Yeah, I got the idea a quite a while ago. ;-)What you're sayin is not what I was asking.Even if I made a long cable with the 31 connectors, I'd still only be able to use 4 key boards and 4 potentiometer boards, right?That's what I was asking.Maybe you can help me there.Greetings for all,Rafa.

hi GeorgeI'm too experiencing the problem with the alps rotaries. I'm using 24 pulse / 24 detents alps rotary encoders via a redec-like board and I always get an increment of two also if 1 is set.Do you know how can I fix it? I've tried setting knitter type but it doesn't work.byep.s.I'm still testing this project by Mike Smits similar to a redec but it allows two encoders with 4 ICs, I still don't know if it is a building error or not but my board works exactly the opposite way as the redec: when you twist the rotary the circuit is opened, while normally it is closed. Do you think it is normal?