Re: [time-nuts] Modern HW replacement for ATOM based NTP server?
On Tue, 07 Apr 2015 13:58:00 -0700, Hal Murray wrote: You can get no-fan style power for Mini-ITX size systems. The ones I'm familiar with have a tiny DC-DC converter that mounts on the big power connector and runs off a laptop size external supply. http://www.mini-box.com/DC-DC I use some of the Wide-Input ATX (Red-ATX plug) supplies from them , they run excellent. Takes up no space , and are totally silent. I'd recommend the WI model , as you can use a spare laptop-psu or whatever you have lying around. CFO ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] Modern HW replacement for ATOM based NTP server?
From: Didier Juges Is the new RPi2 any different in that regard? The RPi 2 has the same basic configuration (Ethernet over USB), but has a 4-core CPU and 1 GB memory, making it very usable as a stand-alone PC. NTP compiles within a reasonable time - about the same as my dual-core Intel Atom. Jitter reported by NTP is just under 4 microseconds, being limited to the sys_jitter value of 3.8 microseconds. Cheers, David -- SatSignal Software - Quality software written to your requirements Web: http://www.satsignal.eu Email: david-tay...@blueyonder.co.uk ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] Modern HW replacement for ATOM based NTP server?
att...@kinali.ch said: The biggest problem would be to get the data into ntp in the right way, as I am not sure whether ntp supports that kind of input. If anybody ever needs help with ntp, feel free to poke me off list. There are 2 ways to get data into ntp. You can implement the kernel PPS API which basically reads a pair of time stamps, one for when the signal changes from low to high and another for high to low. You can feed data in via the shared memory driver. GPSD uses this. Each sample needs 2 time stamps, one for what the time should be and the second for the OS/system time when the first time stamp arrived. -- These are my opinions. I hate spam. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] Modern HW replacement for ATOM based NTP server?
Problematic if you are after microsecond-level accuracy, perhaps, but so would the BeagleBone be. If your needs are more in the 100 microsecond range, either would be fine with a reasonably wide PPS pulse. Not really. If you know how to write C, you can use the timer on the BBB and get to sub-us accuracy levels (IIRC ~10ns). The biggest problem would be to get the data into ntp in the right way, as I am not sure whether ntp supports that kind of input. But I know at least someone working on this. The rpi has, AFAIK, no timer units Attila Kinali = Of course, using special software you can achieve better, but you aren't going to get 10 ns accuracy over Ethernet when using the device as an NTP server, which was the original request. I wonder what accuracy the OP requires? Cheers, David -- SatSignal Software - Quality software written to your requirements Web: http://www.satsignal.eu Email: david-tay...@blueyonder.co.uk ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] Modern HW replacement for ATOM based NTP server?
kb...@n1k.org said: Now you need to sort out the B, the A+ and the B+ in the Raspberry world. There may be more that I have not yet noticed. As far as I can tell, they all are pretty limited once you get past the tight video integration on the B and B+. There is also the 2B with 1 GB and a 4 core CPU. The A, A+, B, and B+ have a 700 MHz CPU. The 2B has a quad core 900 MHz CPU. The A and A+ don't have an Ethernet. The A+ has a smaller card size. The A and A+ have 1 USB port. The B has 2 USB ports. The B+ and 2B have 4 USB ports. The A and A+ have 256 MB. The B and B+ have 512 MB. The 2B has 1 GB. The B+ and 2B have the same connector layout. (they use the same case) http://en.wikipedia.org/wiki/Raspberry_Pi -- These are my opinions. I hate spam. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] Modern HW replacement for ATOM based NTP server?
On Mon, 06 Apr 2015 22:29:23 +, Frank Hughes via time-nuts wrote: Hi,Years ago this forum helped me put together my first GPSDO and NTP server. Using a then-popular INTEL ATOM board, FreeBSD w/ the NTP kernel, 1PPS input from aTB, works great. But as the years go by, HW improves/evolves and it might be time to recreate this functionin some other modern HW. What about the Odroid C1 ? http://tinyurl.com/qd7m4cz Quadcore Armv5 Price as a RasPI Gigabit Enet - Not usb connected CFO ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] Modern HW replacement for ATOM based NTP server?
On Tue, 7 Apr 2015 17:09:28 + (UTC) cfo xne...@luna.dyndns.dk wrote: What about the Odroid C1 ? http://tinyurl.com/qd7m4cz I recommend against using an Odroid. What I have heard from friends is that you do not get any support from the manufacturer in any way. Which means you have to build your own software stack for everything, as there is also no community around these boards. This might work if you have time to tinker, but i guess most of the people here are less into getting an embedded linux system working and more into using such a system as an black box tool. The Cubie borads and the stuff done by Olimex would be also quite good. Especially Olimex is known for their very good user support, as they specifically sell to tinkerers and engineering companies. This also includes that their boards have almost all I/O pins available on connectors. Attila Kinali -- _av500_ phd is easy _av500_ getting dsl is hard ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] Modern HW replacement for ATOM based NTP server?
On 4/8/15 17:13, Attila Kinali wrote: The Cubie borads and the stuff done by Olimex would be also quite good. Especially Olimex is known for their very good user support, as they specifically sell to tinkerers and engineering companies. This also includes that their boards have almost all I/O pins available on connectors. My NTP servers are based around Olimex boards running an RTOS with a custom NTP server (and a PTP client/server as well). Mostly put together for fun. Just left to mount into a case and put some polish on the display... Philip ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] Modern HW replacement for ATOM based NTP server?
The ball bearing fan upgrade is the best idea. Bater idea is to put in a temperature controlled fan so it will spin slow or stop most of the time. I don't see the need to run each server on it's own hardware. Put the cashing DNS server on the same box as the NTP server. Or if you do have two boxes run the NTP server on both In terms of performance, ARM based credit card sized computers do well if you can get the PPS to the general purpose I/O pin that interrupts on an edge. the Pi can't do that the BeagleBone Black can and it sell for $45. It can share a power supply with your Thunderbolt, literally, no fan and it can just steel power from any GPS receiver. The BBB can like run a DNS server at the same time On Mon, Apr 6, 2015 at 3:29 PM, Frank Hughes via time-nuts time-nuts@febo.com wrote: Hi,Years ago this forum helped me put together my first GPSDO and NTP server. Using a then-popular INTEL ATOM board, FreeBSD w/ the NTP kernel, 1PPS input from aTB, works great. But as the years go by, HW improves/evolves and it might be time to recreate this functionin some other modern HW. Looking for a platform not needing a fan. While the ATOM and SSD seem to be OK w/o direct airflow, the Mini ITX Power Supply fan is needed. After three years, the stupid sleeve bearings are beginning to rattlewhich led me to this design review. Recently built a Caching DNS for the home LAN using a Technologic Systems TS-7250-V2, general purpose PC/104 embedded system Runs Debian, 3.14 kernel. Hardware Description:800 MHz CPU (PXA166)512 MB RAMMicro SD card socketRTC with battery and temp compensationTemp Sensor10/100 ethernet port2 USB HS Host ports3x RS-232 serial ports3x TTL serial portsRS-485 portPC/104 connectorUSB device console port (Micro B connector)LCD and DIO Headers75 TTL DIO (including PC/104 conn)5VDC or 8-28VDC power inputA/D converter, 5 channels (with 4-20 mA current loop) A more expensive alternative to the Pi, but reasonable. Wondering if a TS-7250-V2 is worth a try for the NTP function, or if anyone here has already used one of thesefor NTP? Or other non-Pi alternatives? I would very much appreciate some guidance as to the HW/SW combinationthat would best replace the existing NTP server. - 1PPS is available from both a Trimble TB and a Jackson Labs Fury, so I could build a parallel system. Or just put a ball-bearing fan in...73Frank HughesKJ4OLL ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there. -- Chris Albertson Redondo Beach, California ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] Modern HW replacement for ATOM based NTP server?
I think beagle bone black is the answer to this question (because rPi has a USB-network interface that´s problematic) but i´ll let others that know more than me about this specific subject follow on Daniel On 06/04/2015 19:29, Frank Hughes via time-nuts wrote: Hi,Years ago this forum helped me put together my first GPSDO and NTP server. Using a then-popular INTEL ATOM board, FreeBSD w/ the NTP kernel, 1PPS input from aTB, works great. But as the years go by, HW improves/evolves and it might be time to recreate this functionin some other modern HW. Looking for a platform not needing a fan. While the ATOM and SSD seem to be OK w/o direct airflow, the Mini ITX Power Supply fan is needed. After three years, the stupid sleeve bearings are beginning to rattlewhich led me to this design review. Recently built a Caching DNS for the home LAN using a Technologic Systems TS-7250-V2, general purpose PC/104 embedded system Runs Debian, 3.14 kernel. Hardware Description:800 MHz CPU (PXA166)512 MB RAMMicro SD card socketRTC with battery and temp compensationTemp Sensor10/100 ethernet port2 USB HS Host ports3x RS-232 serial ports3x TTL serial portsRS-485 portPC/104 connectorUSB device console port (Micro B connector)LCD and DIO Headers75 TTL DIO (including PC/104 conn)5VDC or 8-28VDC power inputA/D converter, 5 channels (with 4-20 mA current loop) A more expensive alternative to the Pi, but reasonable. Wondering if a TS-7250-V2 is worth a try for the NTP function, or if anyone here has already used one of thesefor NTP? Or other non-Pi alternatives? I would very much appreciate some guidance as to the HW/SW combinationthat would best replace the existing NTP server. - 1PPS is available from both a Trimble TB and a Jackson Labs Fury, so I could build a parallel system. Or just put a ball-bearing fan in...73Frank HughesKJ4OLL ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] Modern HW replacement for ATOM based NTP server?
On Mon, Apr 6, 2015 at 5:29 PM, Frank Hughes via time-nuts time-nuts@febo.com wrote: Looking for a platform not needing a fan. While the ATOM and SSD seem to be OK w/o direct airflow, the Mini ITX Power Supply fan is needed. After three years, the stupid sleeve bearings are beginning to rattlewhich led me to this design review. Recently built a Caching DNS for the home LAN using a Technologic Systems TS-7250-V2, general purpose PC/104 embedded system Runs Debian, 3.14 kernel. Beagle Bone Black. -- Brian Lloyd 706 Flightline Drive Spring Branch, TX 78070 br...@lloyd.aero +1.210.802-8FLY (1.210.802-8359) ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] Modern HW replacement for ATOM based NTP server?
On Mon, 6 Apr 2015 22:29:23 + (UTC) Frank Hughes via time-nuts time-nuts@febo.com wrote: Looking for a platform not needing a fan. While the ATOM and SSD seem to be OK w/o direct airflow, the Mini ITX Power Supply fan is needed. That depends highly on how much knowledge in linux and especially embedded linux you have. If you know how to build your own kernel, what initrd means and today it's actually initramfs and not initrd and you do not fear to build whole software systems by hand, then you can go for one of the many embedded boards. The beaglebone black, the minnow max and SAMA5D come to mind. The Minnow Max has the advantage of being an x86 based system and thus behaves very much like it (not completely though). The BBB has a huge supporting community and you can find a lot of information online. The SAMA5D based systems are nice because Atmel seems to care very much about proper support in the kernel. If you don't feel confortable with the challenges of an embedded system, i would recommend to stick with an ITX (or smaller) PC. Eg The APU[1] from pc-engines, which is the successor of his successful ALIX boards. Pascal Dornier is very helpfull if you have any issues with his boards, and full documentation (including schematics) is online. Attila Kinali [1] http://www.pcengines.ch/apu.htm -- It is upon moral qualities that a society is ultimately founded. All the prosperity and technological sophistication in the world is of no use without that foundation. -- Miss Matheson, The Diamond Age, Neil Stephenson ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] Modern HW replacement for ATOM based NTP server?
albertson.ch...@gmail.com said: In terms of performance, ARM based credit card sized computers do well if you can get the PPS to the general purpose I/O pin that interrupts on an edge. the Pi can't do that the BeagleBone Black can and it sell for $45. What/why can't the Pi do? I have one handy that is processing a PPS on a GPIO pin. David Taylor has a web page with all the fine print on how to set it up. (Thanks.) -- These are my opinions. I hate spam. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] Modern HW replacement for ATOM based NTP server?
Internally the rPI is a ver awkward beast: the CPU is connected to a GPU, and the GPU is connected to the GPIOs... so lots of jitter and latency. It was designed to be a video decoder... the CPU is there for testing and housekeeping. It works, surelly, but it´s not designed to have low latency and jitter. Daniel On 07/04/2015 09:39, Hal Murray wrote: albertson.ch...@gmail.com said: In terms of performance, ARM based credit card sized computers do well if you can get the PPS to the general purpose I/O pin that interrupts on an edge. the Pi can't do that the BeagleBone Black can and it sell for $45. What/why can't the Pi do? I have one handy that is processing a PPS on a GPIO pin. David Taylor has a web page with all the fine print on how to set it up. (Thanks.) ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] Modern HW replacement for ATOM based NTP server?
On Mon, Apr 6, 2015 at 6:29 PM, Frank Hughes via time-nuts time-nuts@febo.com wrote: Or just put a ball-bearing fan in Sites like lilliputing follow the low power/embedded market. I'd suggest starting there. All of my low power systems use power bricks. This includes mini-itx boards. I've had bad experience with a Minnow board and a Sheva. Okay experience with BeagleBoneBlack and RaspberryPi. Good experience with Compulab fit-PC2 systems and Jetway mini-itx fanless boards (the latter two using Atom CPUs). Very good experience with Supermicro fanless mini-itx. It turns out that list is also mostly sorted by price. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] Modern HW replacement for ATOM based NTP server?
I think beagle bone black is the answer to this question (because rPi has a USB-network interface that´s problematic) but i´ll let others that know more than me about this specific subject follow on Daniel == Problematic if you are after microsecond-level accuracy, perhaps, but so would the BeagleBone be. If your needs are more in the 100 microsecond range, either would be fine with a reasonably wide PPS pulse. Yes, the BeagleBone is somewhat better at serving externally according to my own tests, but the Raspberry Pi has the greatest support in the community, if you want to use it for other tasks as well, and doesn't suffer from the problem of generating RF interference at GPS frequencies. Both are excellent little devices! Cheers, David -- SatSignal Software - Quality software written to your requirements Web: http://www.satsignal.eu Email: david-tay...@blueyonder.co.uk ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] Modern HW replacement for ATOM based NTP server?
On Tue, 7 Apr 2015 17:39:01 +0100 David J Taylor david-tay...@blueyonder.co.uk wrote: Problematic if you are after microsecond-level accuracy, perhaps, but so would the BeagleBone be. If your needs are more in the 100 microsecond range, either would be fine with a reasonably wide PPS pulse. Not really. If you know how to write C, you can use the timer on the BBB and get to sub-us accuracy levels (IIRC ~10ns). The biggest problem would be to get the data into ntp in the right way, as I am not sure whether ntp supports that kind of input. But I know at least someone working on this. The rpi has, AFAIK, no timer units Attila Kinali -- _av500_ phd is easy _av500_ getting dsl is hard ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] Modern HW replacement for ATOM based NTP server?
Looking for a platform not needing a fan. While the ATOM and SSD seem to be OK w/o direct airflow, the Mini ITX Power Supply fan is needed. If you are happy with Raspberry Pi or BeagleBone Black, they are the low cost low power way to go. They run Linux. They don't have a real disk. If you do a lot of disk activity, you might wear out the SD card frequently enough to be annoying. That hasn't been a problem for me but I don't have a lot of years on them yet. You could use a USB disk. The Raspberry Pi has a HDMI connector so you can run a real display. The BBB has a mini/micro HDMI connector. I ignore those and ssh in from my main PC. (Some people refer to that mode of operation as headless.) You can get no-fan style power for Mini-ITX size systems. The ones I'm familiar with have a tiny DC-DC converter that mounts on the big power connector and runs off a laptop size external supply. http://www.mini-box.com/DC-DC You could just change the fan or whole power supply whenever the fan dies. Maybe you will get lucky and find one with a fan that lasts a long time. You could try an old laptop. The battery will decay over the years, but it will probably work well enough to cover short dropouts and/or let you wander around for a few minutes. Some laptops have fans. (I have a Dell Latitude D430. It has been running 24x7 for several years. The fan still works. The battery is only good for 10 or 15 minutes.) -- These are my opinions. I hate spam. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] Modern HW replacement for ATOM based NTP server?
Hi! Take a look at https://www.febo.com/pipermail/time-nuts/2014-December/089217.html and https://www.febo.com/pipermail/time-nuts/2014-December/089681.html. There you can see sub-microsecond accuracy. Edésio On Tue, Apr 07, 2015 at 09:52:57PM +0200, Attila Kinali wrote: On Tue, 7 Apr 2015 17:39:01 +0100 David J Taylor david-tay...@blueyonder.co.uk wrote: Problematic if you are after microsecond-level accuracy, perhaps, but so would the BeagleBone be. If your needs are more in the 100 microsecond range, either would be fine with a reasonably wide PPS pulse. Not really. If you know how to write C, you can use the timer on the BBB and get to sub-us accuracy levels (IIRC ~10ns). The biggest problem would be to get the data into ntp in the right way, as I am not sure whether ntp supports that kind of input. But I know at least someone working on this. The rpi has, AFAIK, no timer units Attila Kinali -- _av500_ phd is easy _av500_ getting dsl is hard ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] Modern HW replacement for ATOM based NTP server?
Is the new RPi2 any different in that regard? On April 7, 2015 8:17:12 AM CDT, Daniel Mendes dmend...@gmail.com wrote: Internally the rPI is a ver awkward beast: the CPU is connected to a GPU, and the GPU is connected to the GPIOs... so lots of jitter and latency. It was designed to be a video decoder... the CPU is there for testing and housekeeping. It works, surelly, but it´s not designed to have low latency and jitter. Daniel On 07/04/2015 09:39, Hal Murray wrote: albertson.ch...@gmail.com said: In terms of performance, ARM based credit card sized computers do well if you can get the PPS to the general purpose I/O pin that interrupts on an edge. the Pi can't do that the BeagleBone Black can and it sell for $45. What/why can't the Pi do? I have one handy that is processing a PPS on a GPIO pin. David Taylor has a web page with all the fine print on how to set it up. (Thanks.) ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there. -- Sent from my Motorola Droid Razr HD 4G LTE wireless tracker while I do other things. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] Modern HW replacement for ATOM based NTP server?
Hi Now you need to sort out the B, the A+ and the B+ in the Raspberry world. There may be more that I have not yet noticed. As far as I can tell, they all are pretty limited once you get past the tight video integration on the B and B+. Bob On Apr 7, 2015, at 12:41 PM, Didier Juges shali...@gmail.com wrote: Is the new RPi2 any different in that regard? On April 7, 2015 8:17:12 AM CDT, Daniel Mendes dmend...@gmail.com wrote: Internally the rPI is a ver awkward beast: the CPU is connected to a GPU, and the GPU is connected to the GPIOs... so lots of jitter and latency. It was designed to be a video decoder... the CPU is there for testing and housekeeping. It works, surelly, but it´s not designed to have low latency and jitter. Daniel On 07/04/2015 09:39, Hal Murray wrote: albertson.ch...@gmail.com said: In terms of performance, ARM based credit card sized computers do well if you can get the PPS to the general purpose I/O pin that interrupts on an edge. the Pi can't do that the BeagleBone Black can and it sell for $45. What/why can't the Pi do? I have one handy that is processing a PPS on a GPIO pin. David Taylor has a web page with all the fine print on how to set it up. (Thanks.) ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there. -- Sent from my Motorola Droid Razr HD 4G LTE wireless tracker while I do other things. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there. ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
Re: [time-nuts] Modern HW replacement for ATOM based NTP server?
On 07/04/2015 17:58, Hal Murray wrote: If you are happy with Raspberry Pi or BeagleBone Black, they are the low cost low power way to go. They run Linux. They don't have a real disk. If you do a lot of disk activity, you might wear out the SD card frequently enough to be annoying. That hasn't been a problem for me but I don't have a lot of years on them yet. You could use a USB disk. You can configure the main partition to run read only. This improves lots of things in this regard. Daniel ___ time-nuts mailing list -- time-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.