| Author |
Message |
Dego
Joined: Apr 22, 2008
Posts: 139
Location: Sweden
|
 Posted: Mon Sep 29, 2008 12:27 am Post subject:
|
  |
|
Pin out for .com is right. Thanks for adding that!  |
|
|
Back to top
|
|
 |
chrissugar
Joined: Sep 23, 2008
Posts: 34
Location: ROMANIA
|
| Posted: Mon Sep 29, 2008 12:49 am Post subject:
Re: Sunday News Edition!!! |
|
|
Hi Thomas
| djthomaswhite wrote: |
I am leaving R26 and R32 as is. You can use jumpers there if you really don't like them in the circuit.
|
There is no problem with R26 and R32, they are ok, the only problem is that they are included in the negative feedback network. I did a little hand drawn image. Look at how R25 (R31) should be connected.
Also I added an inverted out as sugested by Serenadi and Bugfight.
| djthomaswhite wrote: |
I could use a second set of eyes in verifying the traces of the schematic versus the board layout. In the attached image, the red color is the copper on the top of the board and the green is the bottom copper. The black are component and silkscreen outlines. We are getting closer to the best possible circuit yes?
|
After a quick look the schematic it seems to look right, but I will check and re-check again.
Also I will try to redraw the schematic from your PCB design, and if it is exactly like the original (plus the correct mods) we can call it is the final version
| djthomaswhite wrote: |
PS. I want to credit Chris Sugar and Serenadi on the PCB for helping with the final design. What are your last names guys? |
My real name is Christian Mike Sugar but on forums I use the chrissugar nickname. 
Thank you Thomas, you are very kind but I don't think I deserve It. No need to mention my name.
chrissugar |
|
|
Back to top
|
|
|
/mr
Joined: Aug 05, 2007
Posts: 223
Location: Elektron City, Sweden
Audio files: 1
|
| Posted: Mon Sep 29, 2008 1:08 am Post subject:
|
|
|
| fonik wrote: | | djthomaswhite wrote: | | Now here is a question... I have the large bottom copper filled plane there. It is not tied to ground. Instead there is a separate ground trace that runs around two edges of the PCB to cover all the necessary ground spots. The question is should the filled plane be tied to ground?If so, can I eliminate the 2-edge ground trace and simply have all ground points run to the filled plane? Or, should I leave the plane unconnected like a piece of shielding that won't serve as an antennae to the ground connections? |
IMO you should leave it as is. a ground plane is good if it has it's own layer (that would increase the cost of the PCB).
if you connected the filled plane to GND you might create closed circles that probably will pic up noise AFAIK. |
I'm no expert on ground planes, but my guess is that it should be connected to ground at ONE spot near the power input. No ground loops that can pick up hum from surrounding fields. If it is left unconnected I would see it as an antenna, not if it is grounded.
But my intuition might be wrong. |
|
|
Back to top
|
|
|
bugfight
Joined: Aug 02, 2007
Posts: 188
Location: Arlington, TX USA
|
| Posted: Mon Sep 29, 2008 9:46 am Post subject:
|
|
|
thanks for the pic, chrissugar, worth a thousand words...
i've read that unconnected copper fill on a pcb is bad,
but don't remember exactly why. gnd planes are good
even without their own layer because they give a good
fat path to gnd.
if you want to open a can of worms, there's the issue of
thermal connections to the gnd plane. (a little cross at
each pad instead of the pad being fully surrounded)
some folks claim they diminish that nice fat connection,
but they sure make soldering easier...
if you don't mind, i'll add one more thing:
some of those pads seem a bit close to each other,
especially around U2. this might be a source of errors
do to bridging from folks like me with shaky hands...
of course, your original sounds awesome and no doubt
we all would have been very happy with that one.
thanks again... |
|
|
Back to top
|
|
|
blue hell
Site Admin
Joined: Apr 03, 2004
Posts: 24445
Location: The Netherlands, Enschede
Audio files: 297
G2 patch files: 320
|
| Posted: Mon Sep 29, 2008 10:11 am Post subject:
|
|
|
Looking at the PDF it seems your vias are almost as large as the holes for the components, there seems to be plenty of copper around it, this should be perfectly ok.
Re. the (ground) plane, when it's not connected to anything you could leave it out just as well, electrically speaking. When you run current through your ground plane you could end up having unexpected behavior, but the same holds for ground traces ... except that it's easier to see with traces where the current will flow. I can't readily see if the whole of the large copper area is interconnected, if it is you could connect it to ground at one single point - the best point likely would be somewhere near the power connector. That way it will add shielding while no current runs through it.
When you are not sure about this you could add two extra holes to allow people to ground the plane with a jumper or leave it out as they please.
_________________
Jan
also .. could someone please turn down the thermostat a bit.
 |
|
|
Back to top
|
|
|
urbanscallywag
Joined: Nov 30, 2007
Posts: 317
Location: sometimes
|
| Posted: Mon Sep 29, 2008 11:09 am Post subject:
|
|
|
| chrissugar wrote: | | There is no problem with R26 and R32, they are ok, the only problem is that they are included in the negative feedback network. I did a little hand drawn image. Look at how R25 (R31) should be connected. |
Here is a conversation between (I believe) Seb Francis and Ken Stone on SDIY, around October of 2006.
| Quote: | >Personally I
>prefer to put the output protection resistor in the feedback loop of the
>output buffer opamp which then gives an effective output impedance close
>to 0ohms unless you try to draw too much current.
Totally agreed, which is how I usually do it.
Ken |
|
|
|
Back to top
|
|
|
Sine
Joined: Sep 10, 2007
Posts: 111
Location: Netherlands
|
| Posted: Mon Sep 29, 2008 11:27 am Post subject:
|
|
|
| Yes that is definalty the way to go for control voltages, it's not that critical for audio signals, but it doesn't hurt to do it anyway. |
|
|
Back to top
|
|
|
bugfight
Joined: Aug 02, 2007
Posts: 188
Location: Arlington, TX USA
|
| Posted: Mon Sep 29, 2008 11:33 am Post subject:
|
|
|
| urbanscallywag wrote: |
Here is a conversation ...
| Quote: | ... gives an effective output impedance close
>to 0ohms unless you try to draw too much current.
... |
|
very interesting. thanks for that.
how is this possible?
isn't the open loop output impedance > 0 ohms? |
|
|
Back to top
|
|
|
/mr
Joined: Aug 05, 2007
Posts: 223
Location: Elektron City, Sweden
Audio files: 1
|
| Posted: Mon Sep 29, 2008 12:31 pm Post subject:
|
|
|
| bugfight wrote: | i've read that unconnected copper fill on a pcb is bad,
but don't remember exactly why. |
I guess it acts like a capacitive connection between all nearby traces, instead of acting like a capacitive ("shielding") connection from every trace to ground. In other words, high frequencies can travel between all kinds of circuit sextions through the unconnected copper. This must be a potential (sic!) source of really bad behaviour!
In addition, I guess it also acts like an antenna - its potential is not fixed and it's free to float around and catch up every possible field and emit it further in bad ways, both fields and the aforementioned capacitive connections to lots of components.
| bugfight wrote: | gnd planes are good
even without their own layer because they give a good
fat path to gnd. |
Yes, if it's used for the ground connections. It doesn't have to be, but it is probably a good idea to skip the separate surrounding ground traces. Important to have the ground connections in a star-like configuration, though.
| bugfight wrote: | if you want to open a can of worms, there's the issue of
thermal connections to the gnd plane. (a little cross at
each pad instead of the pad being fully surrounded)
some folks claim they diminish that nice fat connection,
but they sure make soldering easier... |
This is an important point! When soldering a ground connection that is not thermally disconnected from a plane, you really notice how much more the plane "eats" your applied heat. The solder doesn't melt and fill the pad as easily. Big risk of cold solder joints - and people are likely to get strange boring problems! All solder connections to a plane should have these thermally disconnected solder pads.
 (see the top holes)
Regarding the bad effect on ground connection - it's not really a problem, current is conducted so much better than heat. If you would see the resistance of the arrangement with 4 tiny millimeter-long copper connectors in parallel(!) as a problem, then you would have problems with every millimeter of all your component legs. Cool down!  |
|
|
Back to top
|
|
|
bugfight
Joined: Aug 02, 2007
Posts: 188
Location: Arlington, TX USA
|
| Posted: Mon Sep 29, 2008 12:35 pm Post subject:
|
|
|
found this for calculating output impedance:
http://en.wikiversity.org/wiki/Operational_amplifier#Output_Impedance
for the existing circuit (output resistor in the feedback loop, unity gain, etc),
using this method and assuming (yeah i know an ass out of uming)
an open loop output impedance of 100 ohms, i calc Rout = 1094 ohms,
or slightly less than the suggested configuration (1100 ohms)
but like i said, i'm no expert, there may be more to this story...
why we care about output impedance is maybe even more ot.
but generally: low Zout into high Zin good, napster bad... |
|
|
Back to top
|
|
|
bugfight
Joined: Aug 02, 2007
Posts: 188
Location: Arlington, TX USA
|
| Posted: Mon Sep 29, 2008 12:39 pm Post subject:
|
|
|
| /mr wrote: |
...Regarding the bad effect on ground connection - it's not really a problem, current is conducted so much better than heat. ...
|
totally agree.
sometimes audiophile types claim it raises the noise floor.
until i hear it myself i'll stick to using thermals... |
|
|
Back to top
|
|
|
helitron
Joined: Feb 27, 2008
Posts: 21
Location: vienna
|
| Posted: Mon Sep 29, 2008 2:02 pm Post subject:
|
|
|
looks like I am the last one ...
I'd like to take one PCB if you go for the extended run.
thanks, hel |
|
|
Back to top
|
|
|
urbanscallywag
Joined: Nov 30, 2007
Posts: 317
Location: sometimes
|
| Posted: Mon Sep 29, 2008 3:55 pm Post subject:
|
|
|
| Why isn't R15 in the feedback loop like R26 and R32? |
|
|
Back to top
|
|
|
djthomaswhite
Joined: Nov 22, 2007
Posts: 140
Location: Orange, CA
|
|
|
Back to top
|
|
|
bugfight
Joined: Aug 02, 2007
Posts: 188
Location: Arlington, TX USA
|
| Posted: Mon Sep 29, 2008 9:51 pm Post subject:
|
|
|
i would remove any copper fill that is not connected to gnd,
or move some traces so they are connected.
for example it looks like the upper right area
(under D-SWITCH label) is not... |
|
|
Back to top
|
|
|
chrissugar
Joined: Sep 23, 2008
Posts: 34
Location: ROMANIA
|
| Posted: Mon Sep 29, 2008 11:37 pm Post subject:
|
|
|
| urbanscallywag wrote: | | chrissugar wrote: | | There is no problem with R26 and R32, they are ok, the only problem is that they are included in the negative feedback network. I did a little hand drawn image. Look at how R25 (R31) should be connected. |
Here is a conversation between (I believe) Seb Francis and Ken Stone on SDIY, around October of 2006.
| Quote: | >Personally I
>prefer to put the output protection resistor in the feedback loop of the
>output buffer opamp which then gives an effective output impedance close
>to 0ohms unless you try to draw too much current.
Totally agreed, which is how I usually do it.
Ken |
|
Everything is context in designing with opamps. In this situation it is better to do like I said.
The reduction of output Z is in relation with the value of amplification which in this case is -1 (inverting unity gain circuit).
If you don't know exactly what was the specific circuit in discussion (Seb Francis and Ken Stone ) it is hard to draw conclusions.
chrissugar |
|
|
Back to top
|
|
|
djthomaswhite
Joined: Nov 22, 2007
Posts: 140
Location: Orange, CA
|
Posted: Tue Sep 30, 2008 10:21 am Post subject:
Prototypes have been ordered!
Subject description: Testing to happen shortly |
|
|
Hi gang,
I just ordered the three prototype boards. They will take 1 day for processing and 2 for shipping (business days). Once I receive them I will build and test one myself. The other two are being mailed to co-conspirators for beta testing. Once we have our builds done and tested we will post the findings and share audio/video and pics.
This is an exciting time for me after a couple of weeks of revisions and help from many of you. The summer stage and the placement of the 1K resistor is the only issue to still have some differences in opinion over. The testing folks will see what type of measurable difference, if any, this makes and consider where final placement will be in the production boards.
PAYMENT: I have come to a new plan for production and payment procedures. When the proto is tested, samples and pictures are out there, and the results you can expect from these PCB's are public, then I will contact everyone who has expressed interest in the boards. Here is the plan in steps:
1. I email you to verify you are still interested in boards and how many you want.
2. I send you an invoice for the boards in advance of the production order being placed (see notes why this has changed below)
3. You send me payment via PayPal. Other methods only approved on a case-by-case basis
4. There will be a "pay by" date. (maybe like a 2 week window?)
5. Once payment are accumulated I will have a somewhat accurate count of how many boards to order and the money from you all to do so.
6. Boards will be ordered (10 business day production cycle with 2 days shipping time from my board house)
7. Boards will be received and quality checked, packaged and mailed out to you. Shipping in the US is FREE, International will be discussed back in step 2 invoicing details per your preference.
8. Boards will arrive with you and happiness will ensue! Or, they may go in a drawer for a while like commonly happens with my projects
Okay, here's why the change... There has been great interest in this project. I want to make sure to get boards to as many as who want them without having to cap the run. Pre-pay helps me to do this. Your orders will be combined with my order for some Extra boards. This is necessary for the certainty of future orders and board replacement (i.e. insurance for you that you get your boards).
I have put a lot of time into this project to make sure to "get it right" as they say. From approval from Don to taking design suggestions into consideration I think this will be a great project for all of us in the end! If anyone is uncomfortable with the aforementioned terms I understand... there have been strange times in the past with guys like Mr. R***aleem (QFG) and others for various reasons. I have made different committments than those previous project coordinators. I have a solid eBay seller rating (naturalrhythm) and I have the desire to do this right, as I believe is evident from the start of this whole thing. When boards come in you will be notified here on the board. While I may have to ship about 100 packages it should only take a few days maximum to get your boards out. I will not tell you anything has shipped and you receive it a month later to find it had been postmarked two days before. This would not be in line with what I deem "acceptable business practices." If you are still uncomfortable I can only offer that you will get your boards and the only hang up I can imagine would be getting hung up in customs or similar (although this doesn't happen very often IMOE).
As of right now I only have plans to do a batch ONE TIME! This is due to the change in payment and ordering planning. The project has been posted here on electro-music for a couple of weeks I am I sure anyone interested has said so by now. I am not in the business of making boards and I don't plan to start any regular type of business just yet. So, if you want in say so before the testing is over and payment invoicing begins. If you are uncomfortable with paying in advance send me a PM so I can take your name off the list. There is a small chance I will have extra PCB's and you might be feeling lucky... but you also might miss the boat.
I hope all can respect my changes here. It is close to the Holiday, the the bottom line is I simply cannot order 200-300 PCB's and have people flake on their previously expressed level of interest. I am happy with this plan, my wife (aka banker) is happy with this plan, and it makes good sense for getting enough boards to keep everybody happy. Thanks guys! The boards are at the next step of the process and I hope to have more information to you all SOON! Respectfully,
Thomas
_________________
Thomas
www.naturalrhythmmusic.com/diy.html |
|
|
Back to top
|
|
|
numbertalk
Joined: May 05, 2008
Posts: 992
Location: Austin, TX
Audio files: 5
|
| Posted: Tue Sep 30, 2008 10:39 am Post subject:
|
|
|
| Very exciting. I will gladly pre-pay and am still in for 4 boards - please make sure I'm on the list. Thanks! |
|
|
Back to top
|
|
|
bugfight
Joined: Aug 02, 2007
Posts: 188
Location: Arlington, TX USA
|
| Posted: Tue Sep 30, 2008 11:01 am Post subject:
|
|
|
no problemo on the prepay.
thanks again. |
|
|
Back to top
|
|
|
Dego
Joined: Apr 22, 2008
Posts: 139
Location: Sweden
|
| Posted: Tue Sep 30, 2008 11:06 am Post subject:
|
|
|
| No problem. Keep the wife happy |
|
|
Back to top
|
|
|
frijitz
Joined: May 04, 2007
Posts: 1734
Location: NM USA
Audio files: 54
|
| Posted: Tue Sep 30, 2008 11:08 am Post subject:
|
|
|
| bugfight wrote: | found this for calculating output impedance:
http://en.wikiversity.org/wiki/Operational_amplifier#Output_Impedance
for the existing circuit (output resistor in the feedback loop, unity gain, etc),
using this method and assuming (yeah i know an ass out of uming)
an open loop output impedance of 100 ohms, i calc Rout = 1094 ohms,
or slightly less than the suggested configuration (1100 ohms) |
Ummm ... I don't see how you get that. Perhaps you used the gain of the circuit for A, rather than the open-loop gain of the opamp, which is usually around 10^5. From the link you gave:
| Quote: | | This analysis shows that one of the great benefits of using op amps with large gain is that they make it possible to build circuits with very small output impedance. This means that when connected as a source to other circuit elements, there is a negligible signal loss at the next stage's input. |
Ian |
|
|
Back to top
|
|
|
bugfight
Joined: Aug 02, 2007
Posts: 188
Location: Arlington, TX USA
|
| Posted: Tue Sep 30, 2008 12:03 pm Post subject:
|
|
|
| frijitz wrote: |
... I don't see how you get that. Perhaps you used the gain of the circuit for A, rather than the open-loop gain of the opamp, which is usually around 10^5. ...
|
thanks for looking at this.
yep, that's what i did: used G instead of A...
so with A=10^5, i actually calculate a negative output impedance,
which i suppose means if we pick the right resistors for the opamp,
we could divide by 0 and invent time travel...
(or maybe it's essentially 0 ohms as claimed above,
but i would prefer time travel) |
|
|
Back to top
|
|
|
parasitk
Joined: Sep 21, 2008
Posts: 69
Location: NYC
|
| Posted: Tue Sep 30, 2008 1:22 pm Post subject:
|
|
|
| Yep, no problem on the pre-pay. Very excited about this. Thanks to you and everyone who has contributed for their work on this project! |
|
|
Back to top
|
|
|
djthomaswhite
Joined: Nov 22, 2007
Posts: 140
Location: Orange, CA
|
| Posted: Tue Sep 30, 2008 5:34 pm Post subject:
Forgot to add this |
|
|
For those interested in planning a bit:
Board Dimensions:
3.8 x 2.5 inches each
Spacing between pots is 1.625 inches (just like MOTM vertical spacing). This is from the center of the offset hole to the center of the resonance hole.
PS. The spacing is off on my prototype boards and the posted pdf files of the board so far. This my friends if why we have prototypes I guess. I'll make sure to get it fixed prior to the final run of official boards. I had to chose between 1.625 (Verticle) and 1.720 (Horizontal) MOTM spacing. I chose verticle as it is more likely that multiple units will be used in 1 panel and sideways would make it more difficult to "stack" units to get the classic 4 x LPG configuration.
Thomas
_________________
Thomas
www.naturalrhythmmusic.com/diy.html
Last edited by djthomaswhite on Tue Sep 30, 2008 7:13 pm; edited 1 time in total |
|
|
Back to top
|
|
|
numbertalk
Joined: May 05, 2008
Posts: 992
Location: Austin, TX
Audio files: 5
|
| Posted: Tue Sep 30, 2008 5:37 pm Post subject:
|
|
|
| Is the BOM the same as the one in the zip file you attached several posts back? |
|
|
Back to top
|
|
|
|
Forum index » DIY Hardware and Software
FAQ
Calendar
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
Chat Room
