Have a question about CadSoft EAGLE?  Ask our Expert, Richard! (Archive)

Hi Richard,

frankly, just now I regret that I have upgraded to 6.1., because this version appears much less stable and predictable than 5.11 and there is no way back because of data incompatibility... Just now I found, that when I invoke the gate, which is not on grid in DEV editor, it is placed into SCH with the same offset (off grid), while in 5.11 it was placed correctly on its origin by INVOKE. The ADD command works correctly. It can be fixed in the library, but I think the position of the gate in the DEVICE editor should not have any effect.

I also have lost consistency several times without any visible reason and any warning (only the indicator down right, which I always mentioned only few minutes after the loss occured and wasn't already able to trace it) and had to correct mistakes manually - very annoying and I don't remember similar problem in previous version.

Just now I found, that when I invoke the gate, which is not on grid in  DEV editor, it is placed into SCH with the same offset (off grid), while  in 5.11 it was placed correctly on its origin by INVOKE.

Hi Dusan,

tried what you described, but I can't reproduce it. I changed the 7400 device in the library so that the gates in the device editor are off grid. When I am placing the gates  in the schematic they are placed in the current grid. When I use INVOKE it works the same way. The gate will be placed on the current grid.

It behaves the same as in Version 5.11.

Could you please check this again. If it does not work for you, please send me an example.

Schematic and library file, please.

Best regards,
Richard

Hi Richard,

I have made a SCH to demonstrate the problem: in the LBR I have moved

symbols by 1/2 of standard grid. Try to delete any resistor in SCH and then

invoke back - in this case it seems to work wrongly again, but after the

mouse click the gate snaps to the grid. In my original schematic it didn'

snap and was placed off-grid - and if the offset is small, it  may be seen

as connected properly.. Don't know why, perhaps besause everything was

updated from 5.x, while here in the example everything was done already in

6.1. But also here some problem is visible as the gate is attached to the

cursor with the half-grid offset.

Hi Richard,

design rules are not more saved with BRD in 6.x? Why?? Do you in CADSoft

think it's better to set it after each time BRD is opened again and again or

be aware not to forget to save it in extra file???

I understand, you must be up to eyes in it, but it is the price for early

release of immature product

Dusan

Hi Dusan,

there can occur a problem with the ROUTE command and pads/SMDs that are not on grid. You wrote you don't want to have recommended a beta version, but this is exactly what I should do now. The current beta 6.1.1 has this problem fixed. If you don't want to use 6.1.1 you could also try aworkaround in version 6.1.0. When you activate ROUTE and draw a short wire, drop the wire from the mouse (with Escape) and and go on routing again. Don't activate ROUTE again (it is still active), simply click onto the end of the wire and go on. Now the snap problem should be gone.

Regards,
Richard

Hi Richard,

do you have any idea why, though the Snap length is set to 0.5mm, 6.1 leaves

tens of short airvires (0.06mm long and so) and do not connect it to the

pads??? Must be connected manually and I have to click lot of times trying

to hit the pad as exacly as possible, making unwanted bundles at the ends of

wires. What's interesting, sometimes Snap obviously works, sometimes it is

nearly impossible to finish the connection. Oh my God, have anybody tested

6.x before it was released??? Really horrible "upgrade"...

And please, do not recommend me to try beta version, I need to go forward,

not to test for free and on hot data the software I have purchased. I am now

working in both 6.1 and 5.11 and it is really a big relief to go back to

5.11, but it is quite uncomfortable to keep all data separated and not

damage anything with 6.1.

I think CadSoft have lost a lot of its good name, pity.

Dusan

Hi Richard,

I think SNAP doesn't work, at least with SMDs - difficult to hit the SMD

center even with large Snap length

Dusan

Hi Dusan,

please tru to open, than make ERC (OK), pinswap any port connections .................and repaet ERC (again OK) . Now UNDO and again repeat ERC - the consistency is lost and connections

mismatched in SCH

I see the problem.... forwarded to our development team.  Thanks,

Richard

Hi Dusan,

Another minor bug: if I use PINSWAP, the first selected pin is not highlited at its real position, but in  the other part of drawing relative to (0 0) instead to the device origin. Minor bug, but I think there are still too much similar bugs in 6.1, lot fo them much serious (primarily the sudden consistency losses are quite annoying)

this problem is already reported to our development team. We will fix it as soon as possible.

Sorry for wasting your time, perhaps some as bug report thread would be worthy in the forum.

It's no waist of time. Each problem reported helps us to imporove our software.

Regards,

Richard

Hi Richard,

I have made a sample file:

please tru to open, than make ERC (OK), pinswap any port connections (say

P2.0 - P2.1) (the swaplevel was temporarily changed on all ports to 1 to

obtain a possibility to optimize layout) and repaet ERC (again OK) . Now

UNDO and again repeat ERC - the consistency is lost and connections

mismatched in SCH - can be only fixed in XML files or by deleting IC1 and

placing again (i naven't found other way). Try to SHOW swapped-unswapped

nets - it really doesn¨t look as correct behaviour!

Dusan

Hi Richard,

thank you for the information.

Yes, I agree that XML gives some advantages (I also have fixes some of the

consistency losses direcly in XML) and that EAGLE is (or was?)very modest as

to the memory/disk apace requirements, but though I would appreciate the

option binary/XML - but this is not a problem for me.

If I am reverting to 5.11 now, it is not forever, but only until the bugs

will be fixed - world is going forward and one can't stay on place. I think

6.1 is not mature yet and if you are scheduled, it is better to stay in 5.x.

Another minor bug: if I use PINSWAP, the first selected pin is not highlited

at its real position, but in  the other part of drawing relative to (0 0)

instead to the device origin. Minor bug, but I think there are still too

much similar bugs in 6.1, lot fo them much serious (primarily the sudden

consistency losses are quite annoying)

Sorry for wasting your time, perhaps some as bug report thread would be

worthy in the forum.

Hi Dusan,

you wrote:

I copy elements using COPY it is quite difficult to place part correctly - it tends to snap to the neighbouring grid position (in all directions), so sometimes I have to undo and try again several times until I place element correctly.

This is a problem which is already fixed in version 6.1.1, the current beta.

BTW, I think that it would be great to have the possibility to back-convert the data or to have the option to save binary too (also because of much smaller files).

The XML data structure offers lot of benefits which are not at once visible. But I am sure that XML gives big incentive for the future. We feared that XML data would be ten times more than the binary data, but it is "only" about 5 times more. Depending on the contents, of course. Today we are in the lucky situation that it is no longer so important whether you have a 200k file or 1Mbyte. Compared to other applications EAGLE is very modest.

Talking about reverting to the previous version: You could try two ULPs: export-board.ulp and export-schematic.ulp from the cadsoft.de download area. This might work.

Regards,

Richard

Hello Richard,

i think that there is a genereal problem in 6.1 with placing parts - also if

I copy elements using COPY it is quite difficult to place part correctly -

it tends to snap to the neighbouring grid position (in all directions), so

sometimes I have to undo and try again several times until I place element

correctly. 

Currently I am considering going back to 5.11, because I think 6.1 is still

not mature, but because od data incompatibility it would mean loss of a lot

of work - BTW, I think that it would be great to have the possibility to

back-convert the data or to have the option to save binary too (also because

of much smaller files).

Hello Richard,

i think that there is a genereal problem in 6.1 with placing parts - also if

I copy elements using COPY it is quite difficult to place part correctly -

it tends to snap to the neighbouring grid position (in all directions), so

sometimes I have to undo and try again several times until I place element

correctly. 

Currently I am considering going back to 5.11, because I think 6.1 is still

not mature, but because od data incompatibility it would mean loss of a lot

of work - BTW, I think that it would be great to have the possibility to

back-convert the data or to have the option to save binary too (also because

of much smaller files).

Hi Richard,

design rules are not more saved with BRD in 6.x? Why?? Do you in CADSoft

think it's better to set it after each time BRD is opened again and again or

be aware not to forget to save it in extra file???

I understand, you must be up to eyes in it, but it is the price for early

release of immature product

Dusan

Hi Dusan,

there can occur a problem with the ROUTE command and pads/SMDs that are not on grid. You wrote you don't want to have recommended a beta version, but this is exactly what I should do now. The current beta 6.1.1 has this problem fixed. If you don't want to use 6.1.1 you could also try aworkaround in version 6.1.0. When you activate ROUTE and draw a short wire, drop the wire from the mouse (with Escape) and and go on routing again. Don't activate ROUTE again (it is still active), simply click onto the end of the wire and go on. Now the snap problem should be gone.

Regards,
Richard

Hi Richard,

do you have any idea why, though the Snap length is set to 0.5mm, 6.1 leaves

tens of short airvires (0.06mm long and so) and do not connect it to the

pads??? Must be connected manually and I have to click lot of times trying

to hit the pad as exacly as possible, making unwanted bundles at the ends of

wires. What's interesting, sometimes Snap obviously works, sometimes it is

nearly impossible to finish the connection. Oh my God, have anybody tested

6.x before it was released??? Really horrible "upgrade"...

And please, do not recommend me to try beta version, I need to go forward,

not to test for free and on hot data the software I have purchased. I am now

working in both 6.1 and 5.11 and it is really a big relief to go back to

5.11, but it is quite uncomfortable to keep all data separated and not

damage anything with 6.1.

I think CadSoft have lost a lot of its good name, pity.

Dusan

Hi Richard,

I think SNAP doesn't work, at least with SMDs - difficult to hit the SMD

center even with large Snap length

Dusan

Hi Dusan,

please tru to open, than make ERC (OK), pinswap any port connections .................and repaet ERC (again OK) . Now UNDO and again repeat ERC - the consistency is lost and connections

mismatched in SCH

I see the problem.... forwarded to our development team.  Thanks,

Richard

Hi Dusan,

Another minor bug: if I use PINSWAP, the first selected pin is not highlited at its real position, but in  the other part of drawing relative to (0 0) instead to the device origin. Minor bug, but I think there are still too much similar bugs in 6.1, lot fo them much serious (primarily the sudden consistency losses are quite annoying)

this problem is already reported to our development team. We will fix it as soon as possible.

Sorry for wasting your time, perhaps some as bug report thread would be worthy in the forum.

It's no waist of time. Each problem reported helps us to imporove our software.

Regards,

Richard

Hi Richard,

I have made a sample file:

please tru to open, than make ERC (OK), pinswap any port connections (say

P2.0 - P2.1) (the swaplevel was temporarily changed on all ports to 1 to

obtain a possibility to optimize layout) and repaet ERC (again OK) . Now

UNDO and again repeat ERC - the consistency is lost and connections

mismatched in SCH - can be only fixed in XML files or by deleting IC1 and

placing again (i naven't found other way). Try to SHOW swapped-unswapped

nets - it really doesn¨t look as correct behaviour!

Dusan

Hi Richard,

thank you for the information.

Yes, I agree that XML gives some advantages (I also have fixes some of the

consistency losses direcly in XML) and that EAGLE is (or was?)very modest as

to the memory/disk apace requirements, but though I would appreciate the

option binary/XML - but this is not a problem for me.

If I am reverting to 5.11 now, it is not forever, but only until the bugs

will be fixed - world is going forward and one can't stay on place. I think

6.1 is not mature yet and if you are scheduled, it is better to stay in 5.x.

Another minor bug: if I use PINSWAP, the first selected pin is not highlited

at its real position, but in  the other part of drawing relative to (0 0)

instead to the device origin. Minor bug, but I think there are still too

much similar bugs in 6.1, lot fo them much serious (primarily the sudden

consistency losses are quite annoying)

Sorry for wasting your time, perhaps some as bug report thread would be

worthy in the forum.

Hi Dusan,

you wrote:

I copy elements using COPY it is quite difficult to place part correctly - it tends to snap to the neighbouring grid position (in all directions), so sometimes I have to undo and try again several times until I place element correctly.

This is a problem which is already fixed in version 6.1.1, the current beta.

BTW, I think that it would be great to have the possibility to back-convert the data or to have the option to save binary too (also because of much smaller files).

The XML data structure offers lot of benefits which are not at once visible. But I am sure that XML gives big incentive for the future. We feared that XML data would be ten times more than the binary data, but it is "only" about 5 times more. Depending on the contents, of course. Today we are in the lucky situation that it is no longer so important whether you have a 200k file or 1Mbyte. Compared to other applications EAGLE is very modest.

Talking about reverting to the previous version: You could try two ULPs: export-board.ulp and export-schematic.ulp from the cadsoft.de download area. This might work.

Regards,

Richard