We don’t need Openlane and Caravel. If you can jus...
# analog-designa
Amro Tork
12/02/2020, 4:59 PMWe don’t need Openlane and Caravel. If you can just provide us with a GDS with sealring and a couple pads if possible??
k
Kunal
12/02/2020, 5:05 PMWe are stuck with PNR with pads here
https://github.com/rsnkhatri3/vsdPLLSoC
a
Amro Tork
12/02/2020, 5:06 PMI prefer to have full control over the chip.
Amro Tork
12/02/2020, 5:07 PMIf you ate trying to target High frequency design no way OpenLane could help
k
Kunal
12/02/2020, 5:15 PM100MHz max? Is this high? For 130nm, I suppose it should be around 130-140MHz
c
Chris Jones
12/03/2020, 10:12 AM@Kunal I've done 4GHz VCOs on 0.35um CMOS (with divider in bipolar transistors) and 4GHz VCOs and counters / dividers on 0.18um pure CMOS.
a
Amro Tork
12/03/2020, 10:12 AMI was going to say that it's possible to get something in that range.
Amro Tork
12/03/2020, 10:12 AM@Chris Jones
Amro Tork
12/03/2020, 10:12 AMAre you taping out something in that range?
Amro Tork
12/03/2020, 10:13 AMIf yes, let's connect privately, I have built an Inductor library at my end.
Amro Tork
12/03/2020, 10:15 AMOne question @Chris Jones I was hoping to tapeout something that work in the ISM band 2.4GHz range. I think this might not work given the Openlane requirement.
Amro Tork
12/03/2020, 10:15 AM?
Amro Tork
12/03/2020, 10:15 AMWhat do you think?
c
Chris Jones
12/03/2020, 10:15 AMNo, I have given up on this tapeout, because I assume that I have missed the deadline, and family / work commitments mean that I am unlikely to finish anything before it really does go to fab. Also I find the learning curve of magic very steep so that my progress was very slow, so might wait for more support for this process in KLayout etc.
a
Amro Tork
12/03/2020, 10:16 AMMy team is working actively on creating device generators for Klayout
Amro Tork
12/03/2020, 10:17 AMWe have both NFET , PFET and Inductors so far.
Amro Tork
12/03/2020, 10:17 AMonly 1.8V
Amro Tork
12/03/2020, 10:17 AMNo RF for now.
Amro Tork
12/03/2020, 10:17 AMIf I have known about this tapeout a couple of month back we would have made a difference.
Amro Tork
12/03/2020, 10:18 AMThe problem with inductors is that this process is AL
Amro Tork
12/03/2020, 10:18 AMand I think they would be too lossy.
Amro Tork
12/03/2020, 10:18 AMFor any real applications.
c
Chris Jones
12/03/2020, 10:20 AMA lot of cellphone chips were made with Al inductors. The ones I did were 3um aluminium. It is not at all useless. Sure RDL would probably be better but that has other issues.
a
Amro Tork
12/03/2020, 10:21 AMyeah, but losses are really high.
Amro Tork
12/03/2020, 10:21 AMYou have to make sure the LNA is doing a great job and on both SNR and gain.
c
Chris Jones
12/03/2020, 10:26 AMIt depends on your application. For example, this process would be good enough to make a cellphone transceiver for 3G, including the LNAs. Sure there are things that it could not do, maybe the trick is to identify applicaitons that can be done in this process - there are plenty.
a
Amro Tork
12/03/2020, 10:27 AMK is really high for the layers dielectric layers
Amro Tork
12/03/2020, 10:29 AMat the top layer
Amro Tork
12/03/2020, 10:29 AMAmro Tork
12/03/2020, 10:29 AMI agree with you.
2 Views