Comparison Between Protel P99SE and Protel DXP
This is currently an unformatted dump of some
email posts (mainly from the PEDA and Altium DXP forums). Maybe one
day I will organise this more elegantly. This page was
updated after the release of Protel 2004 - it does not really go
into any detail of the Nexar stuff, just the PCB, Sch, Sim sort of
tools. DXP 24004 was released a long time ago. This comparison does not discuss features, improvements, problems, bugs etc from later versions. This document is no longer very useful and is really just a piece of history. Maybe some P99SE users will find some value in it but please don't take this as remotely representing the current state of the Altium software.
Here is my summary, very incomplete,
very personal and in no particular order.
Bi-Di blowing and sucking from P99SE seems OK (with some provisos on the use of names
for parameters etc, and some new rules can't convert of course)(There are some recent reports of some issues - check the DXP forum
Speed on my old PIII-450 256MB is marginally acceptable (I did not really expect this). There is a significant speed hit compared to P99SE, though. (Note DXP only currently works under Win2k and XP, not NT or 9x.).
I guess P2004 would be slower than DXP and so be no go on the old
Integrated libraries offer a better level of control in some circumstances - library contains sch symbols, footprints, sim models etc.
Stability generally seems OK to me. I have had a number of AV's but only once lost any data, and that was doing a non-core operation (copying and pasting from a sim chart into a word processor, this may have been fixed) - I am not sure what others would say about stability but I do not see a lot of complaints on the DXP forum. Update for P2004 - I have recently had one crash that required killing DXP without first being able to save. See Dislikes Point
Ctrl-R rubber stamp placement mode (like pasting but no need to keep Ctrl-V'ing).
I quite like the new panel based view - though dual monitors make a huge difference. The Inspector panel is good. (I remain unimpressed by the merging of the List view with the Query/Filter entry panels, see here for details.)
I like the query system - though I understand some find it difficult. The "Find Similar Objects" right-click function is good.
DDB has gone to make way for disk files that
work well with version control systems. (See also DXP Dislikes point 9.)
Interface to version control systems.
But here is the biggy, the killer feature, the solution to all our ills and the bringer of world peace - the little-arrowy-thingy-next-to-the-file-menu has a text, yes a text, label!
Scripting now has the power to access the underlying CAD data and iterate over components, tracks etc. No need to have a copy of Delphi (and the right version at that) to develop complex add-ons.
Sch and Sim
The new query engine is great. It is more complex and possibly does help if you have some experience with software development and booleans. It is much more powerful than the old global operations stuff. The applies to Sch and PCB.
The ability to use version control properly is a big plus.
Right-click and drag panning in Sch. I now hate it when I go back to P99SE and try editing schematics.
Freely addable parameters at both the library and Sch level. (Pins and projects can also have parameters.)
Better control of footprint source - you can choose which library, or let it choose first available and quite a few options along these lines.
Much better ERC - including the important ability to check for unplaced parts of a component (power supply sections, preventing floating inputs...). P2004 - ERC is on-line - red wiggles show
something is wrong. I find it quite useful - traps errors early.
The initial problems with the loss of selected vs focussed are now watered down by provision of a number of tricks, and the provision of a selection memory, such that this is no longer an issue for me - this was one of my *major* reasons that I could not go with DXP seriously. I am not sure the new system is as good as the old, but it is certainly no longer a critical issue at all, for me at least.
Database linking is much improved (including to Access and SQL-based including MySQL etc) -
and no longer so terribly slow.
<comment on feature
removed due to DXP SP2 Beta>(I forgot the feature I wrote about here originally was affected by the latest SP Beta. Take it as read that this is a very nice and powerful feature that extends greatly
P99SE abilities. Couldn't be bothered re-ordering all the list numbers.)(SP2 has been released but now I can't recall what this was to add it back in - oh well.)
Sch Part field can be locked to reduce the chance of accidental changes - useful for those that work with a "one-symbol per company part number" - a 4k7 resistor will not have its value entered instead of replacing the part from the library. Though a field locked in the SchLib can still be unlocked in Sch - I would like to be able to *fully* lock a parameter within the SchLib to prevent any changes in the
Filtering and masking features - stuff can be masked off (greyed out) and then not subject to edits. You would have to see it to understand it. Linked to the query system.
Dropping a two pin component onto a wire will (optionally) break the wire.
Co-linear wire segments are (optionally) joined - so removing stray autojunction hotspots.
Much better dual monitor support.
Support for multi-channel designs, including some user suggested options for naming components in the channels.
Better support for build variants - though I have not tested this in an serious fashion. Design variants is something we do heaps of.
Much better simulation viewing.
Easier mechanism to integrate sim (and other) models into a component. Though there are still some quirks here, nothing like as convoluted as P99SE though.
Having our user suggestions (if presented well with a clear justification) implemented a few weeks later - get involved and your worthy ideas get implemented. It is great to see in a program that many will not easily be able to change from - the pain level in changing a CAE pkg is high. I take the view that it is much cheaper for me to get involved and see my ideas appear than it is to move CAE vendors -
at least up to some bug/misfeature level anyway.
Control over hidden pin connectivity on a per pin basis.
Sch footprint models can have a user defined pin map. So a Sch part may use pin number, 1, 2, 3 etc, while the PCB footprint may be a BGA or similar with a row, column pin/pad/ball naming scheme (such as A1, A2, ..B1, B2...). In DXP Sch (and SchLib) you can establish a pin map between the different pin naming scheme and PCB naming scheme. This is very useful when a part is available in both TQFP or BGA footprints - one Sch part can do both footprints, if desired.
Much more powerful rule system.
Board shape stuff is OK. If it helps split plane checking, then I can live with it. It may
even prove to be a winner. It certainly means that you can have mechanical details and sheet templates etc and still get reasonable reports from board info.
Auto-pan issues fixed - hopefully forever but I certainly can't tell for sure on my old clunker. (Recent note 26Sep03 - auto-pan seems fine on my AMD 2700+/ATI Radeon 9700 based machine. BTW I haven't had any issues with this ATI graphics card under DXP (or P99SE).)
Intelligent dimensions that can stay associated (and rescale) with entities.
A much wider array of dimensioning tools (radius, angle, baseline etc).
Net ties - should be able to do away with most of those kludges like wiring on unused layers and the "Lomax Vitual Short"
Flipping selections works correctly - no need for multiple pass operations as required under P99SE. (Recent note 26Sep03 - though there is a problem, currently, with this - see PCB point 13 for details.
Much better rules for checking component (and object in general) locations (InRegionAbsolute, InRegionRelative, TouchesRoom, WithinRoom queries). DRC is much better
Ability to copy room
formats (multi-channel design). (There is debate on whether
this feature could be improved - mainly should tracks that overlap
the room boundary, even by tiny amounts, be copied - currently
there tracks that cross the room boundary are not copied - meaning
possibly some cleanup for tracks feeding in and out of the
channels. I think that there are some improvements to be had
PCB components have
types as well as Sch components - to prevent them from being
affected during updates from Sch - stuff like mounting holes or
fiducial components. Also, Net ties are implemented using the
component type feature. P2004 added the missing type - a component
that should be updated on the PCB but not added to the BOM (example
non-net tie RF PCB-only components).
Once (recent note
26Sep03 - some say "if") it is working, the autorouter has much
greater control over weights and costs.
Mech layers can be associated (Layer
pairs) so they flip layers when components etc are swapped from top
to bottom etc. Much like top and bottom solder masks are
associated. (Recent note 26Sep03 - there is a bug with the updating
a PCB footprint from a library for a component that has been
flipped, and uses the paired layers - currently paired mech layers
are not massaged if the target component has been flipped.
Hopefully this will be fixed soon.)
PCB polygon edit tools
are a lot better. (Tony K. reminded me of this).
issues when swapping between metric and imperial units seem to have
been fixed or at least greatly improved. But See Dislikes point 8.)
Filtering and masking
(see Sch point 11). P2004 adds
highlight modes which is good for inspecting nets on the PCB.
Also, P2004 adds better support for masking for users that prefer
white backgrounds for PCB - so I believe anyway.
still various Access Violations. Sometimes these force you to
crash the program (using Task Manager). Sometimes I find they
appear and then go away and all seems to be OK - though I usually
re-start. P2004 is not as crash free as P99SE.
The autorouter still
needs lots more work, it is not yet the breakthrough many had
hoped. There is lots of work going on on this at the moment I
gather. This would be my biggest disappointment. I still see poor
autorouting following the release of P2004 - clean up is much
better and completion is better but it is much slower than the
P99SE router but it does do better routing. Still not a
breakthrough from what I can see.
The UI is different and
a number of the dialogs are harder to read than P99SE. They may be
useful for new users but one rapidly looses this benefit as you
become familiar. This would not be a go/no-go factor in a buy
decision but it is an issue that is coming up in discussions at
times. The eye candy is of no importance to me - I'd prefer
concentration on an efficient user interface and if this improves
the look then great. I don't like eye candy that makes it
harder to find the stuff I want in dialogs - there are some
instances of this.
As mentioned the loss of
selected vs focussed was originally a major issue. Chnages
have been made to overcome the loss of functionality. This would
only be a minor issue for me now as I can certainly work
effectively with the updated editing mechanisms and tools. Shame it
was lost though.
Forced upgrade to
Win2k/WinXP. I was already using Win2k so it is not a big issue for
me, but I hate to see programs encouraging users to pay MS more
Slower than P99SE - but
what is surprising about this. My main beef here would be that some
(much?) of the speed loss is possibly due to eye candy rather than
features. Still, as has been discussed before, for most engineers a
1 to 3 year lifetime for a computer system is not atypical. I just
hate it when caught by the out-of-sync upgrade cycles (like I am
There are still bugs, of
course. There are also some improvements that users are wishing
that have not (yet?) been implemented. But I, and others, are very
impressed with the speed of progress and as I said it is great to
be able to influence the product. Those that are not involved will
have to stay with P99SE, move to something else, or "suffer" what
about a dozen other users convince Altium is worthwhile. (Recent
note 26Sep03 - the rate of visible progress on changes has slowed
significantly, with Service Pack 3 taking a *long* time to mature
and be release publicly. This is disappointing.)
underlying base units have not been metrified which could, if
properly done, have resolved all rounding issues.
gone - but see General point 8.
The DDB had some nice characteristics, but nothing really that
can't be dealt with with disk folders and a good compression
Still no generalised (not just
components or electrical entities) clearance rule with overlap
capability (ability to set negative clearances). This feature
is not available in P99SE, either, but is an often requested
feature. Maybe in a future service pack.
Loading external netlists is harder
than it should be. Some users have reported problems but I
have not seen this and I even though I have asked for some examples
to try out I have not received any.
The Parameter Manager is good but needs
more work to allow us to deal with so-called "system
parameters". System parameters can't be added in Parameter
Manager, only changed. This is a nuisance.
In PCB, when you click on copper, the
PCB panel does not update and show the nodes connected to
this net. P99SE did this; users have asked for this feature to come
back. It could hook in with the Ctrl-Click highlight
mode. Irritating, it will be very annoying if it doesn't come
Some experienced P99SE users are having
trouble switching to the new method of doing globals. They
report it is harder for them to remember the keywords. They miss
the always available nature of the P99SE globals. There are
certainly some operations that are much more fiddly in DXP/P2004
than in P99SE - changing the attributes of designators or comments
of selected components is one. There is a change that could
be done to fix this but it has not been implemented (hopefully
Tony Karavidas made these comments on the PEDA
forum - I hope he doesn't mind me quoting him
Placing sch pins now allows
incrementing/decrementing pin numbers WHILE decrementing/incrementing pin names. I've been wanting that one
for along time!!
Right mouse button pan in SCH and PCB.
Mechanical component type in SCH...keeps some
things from ending up on the PCB but still
remain in BOM.
Very useful highlighting/masking features. It
really helps to trace things around the
schematic and the PCB.
SCH net 'Graph' mode which shows (very clearly)
where a net connects using this nice looking
floating graphic line and it works in a complex hierarchy.
Navigator panel is very useful and far beyond the
old Browse panel.
Error messages all have live links to the design
showing the errors.
PCB polygon edit tools are a lot
PCB dimensioning tools are a LOT
for thinking the query user interface should not be on the
- The List panel is a grid-based view of the
- It is most sensibly laid out as wide and short
(across the bottom of the screen), and generally hidden, in my
- The data in the List panel is only used
- Showing the List panel can cause significant
slow downs in DXP as all the grid data is updated (grid controls
are often slow). So keeping the List panel hidden as much as
possible is desirable.
- Query entry is much more commonly used than the
List panel data. It is desirable to have the query UI available at
all times or on a readily available pop-up (probably arranged of
the side of the workspace or on the secondary monitor).
- The query entry user interface is better laid
out as taller, and need not be as wide, in my opinion.
- These conflicting optimal layouts, the fact that
filling the List with design data can be a slow operation and the
differences in the amount of use the query UI gets verses the List
panel, suggest that the query UI should be be split from the List
- My view is that the Query UI and the Inspector
could be merged, or go back to the original arrangement of a
separate Query panel.