In my last round of testing I found that Unraid v6.8 SMB still underperforms compared to Windows Server 2019, but I was wondering if it is a Linux Samba problem, or an Unraid problem.
I installed an Ubuntu Server 18.04.3 LTS VM on Unraid, bridged network, 16GB RAM, 128GB raw disk located on the BTRS cache volume, consisting of 4 x Samsung Pro 860 SSD drives. This is exactly the same configuration I use for the W2K19 test VM. I installed Samba on Ubuntu using default options.
I created a SMB share that is backed by the VM disk image, and a second share that is mapped directly to an Unraid share located on the cache volume. For both shares the Ubuntu VM and Samba server will handle SMB network traffic, but one share will write to the Ubuntu EXT4 volume backed by the VM disk image, and the second will write through to the underlying Unraid BTRFS cache volume using VirtFS.
I ran a series of tests using my DiskSpeedTest utility, and the results are below.
Note that the VirtFS mapped share exhibit some problems that appear to be caching related. E.g. the file iteration test would create 14000 files, but iterating the just created files would only read 3080.
My conclusion is that the Linux Samba SMB performance is on par with that of Windows Server 2019, and that the performance problems are attributed to the Unraid file write performance. The Windows test used NTFS and Ubuntu used EXT4, so it could be BTRFS and XFS related, but more likely something Unraid does. Maybe the next step could be to test a bare metal Ubuntu SMB on XFS and BTRFS.
In a previous post I wrote about my transition from Nest to Ecobee3 thermostats, and how the biggest benefit of the E3 was the use of remote sensors.
After several months of use, winter and summer, I find the remote sensors really do work very well, and our bedrooms remain at the desired temperature, while the areas around the thermostats can be warmer or colder.
But, the E3 is not perfect, there are two recurring problems; the remote sensors would report offline, and the units would lose network connectivity.
I’ve received sensor offline alerts a couple of times, typically happens early mornings, maybe interference, don’t know, the sensors never move from where they are placed.
Every time a I get a sensors offline report, the sensor already restored connectivity. This behavior is very annoying, I will get two emails a minute apart, and the E3 UI will have an alert saying sensor offline, I click ok, and then immediately an alert saying sensor online.
I expect the E3 to have some sort of grace period before it deems a problem so important that it needs to notify me. As is, it is just an annoyance as there is no remediation action to take.
The second problem is the E3 loses network connectivity, this is a real problem, as the units remain offline until power cycled, and to power cycle the E3 has to be removed from the wall bracket, i.e. there is no reboot menu option.
I reported this problem to Ecobee support in October, and on the SmartHomeHub community forum. Yes, it is a bit pathetic that Ecobee does not have their own support community forum. Ecobee had me reserve static IP’s in the DHCP server, setup a dedicated 2.4GHz SSID, still disconnects. By December the problem was still happening, and Ecobee support escalated the problem to their development team, it is two months later, and still no updates from Ecobee support on the problem.
Through my own research and experimentation I suspect the problem to be that the E3 is unable to handle a WiFi channel change, and unable to roam between access points. There are some conditions that trigger the problem that I cannot explain.
I have multiple access points in my house, same SSID, different channels, 2.4GHz and 5GHz bands. I tested with Ubiquity UniFi AC and with Ruckus/Xclaim Xi-3. And in case you’re wondering, no other devices in my house have any problems with WiFi, even with dynamic channel selection.
I can make the E3 fail by either changing the AP channel, or by making it roam to a different AP, also changing channels. If I configure the AP’s to use auto channel selection, then the E3 will fail as soon as the AP chooses to change channels (UniFi does this on startup, Xclaim does this dynamically). If I manually change the AP channel, the E3 will fail. If I take one AP offline, the E3 will fail to roam to a different AP (on a different channel).
Even with my AP’s configured with static non-overlapping channels, the E3 would still sometimes fail, requiring a power cycle. I do not know why this would happen, as signal strength by the E3’s are perfect.
On the plus side, the E3 units remember the schedule, and even when offline, they continue to operate.
Bottom line is E3 WiFi is not reliable, and E3 support/dev is not responsive.
This post is about my research into finding suitable MR16 LED’s for replacing the 50W recessed halogen lights in our house. In summary, I’ve found “ok” bulbs, not great bulbs, and you can read about the details below.
Our house is about 3 years old, new construction, and one of the many decisions we made during planning was recessed halogen vs. recessed LED lighting. At the time my calculations showed the additional cost for LED lights would only be recovered in electricity cost savings after about 12 years, not worth the cost at the time. Another problem was the optical quality of the products, the near halogen optical quality LED products were ridiculously expensive, and the mainstream LED’s were of poor optical quality, and had poor dimmability.
Given the situation we opted for recessed Elco MR16 low voltage 12V AC magnetic transformer halogens, and planned on retrofitting them with LED’s as the technology improved and costs came down.
Now, 3 years later; our electricity cost is way higher than originally estimated, we installed solar that gave us a 50% reduction in cost, some of the recessed reflectors are showing signs of heat damage from the halogen bulbs, and 12V MR16 LED’s have entered the mainstream.
I’ve been looking for MR16 LED’s for some time now, same problem as 3 years ago, dimmable good optical quality bulbs are very expensive, ~$20 per, while eBay and Amazon sourced Chinese manufactured no-name brands are ~$4.
During my research I’ve made a few important observations:
US electrical code requires the use of GU10 bi-pin twist-lock lamp bases for new construction, and GU10 will eventually replace all E26 style screw in bases. In the past months I found that there is a much wider supply of 110V GU10 base MR16 dimmable LED bulbs compared to GU5.3 12V bulbs. This is especially true for the no-name brand Chinese suppliers on eBay. I am assuming that the electronic circuitry used is similar to that used in the widely available regular E26 / A21 110V dimmable LED bulbs, and that the only difference is the MR16 housing construction. Keeping in mind that most installed AC dimmers are forward phase, and support a large variety of load types, while an LED is a constant current device that typically uses pulse width modulation for dimming. Thus a line voltage forward phase dimmer to pulse width modulated LED driver circuit is non-trivial, adding a 12V AC transformer in the mix, and supporting both 12V AC and 12V DC loads further complicates the circuitry, especially when dimming is required.
Just like there is an expanding variety of dimmable GU10 line voltage MR16’s, there is an equal growing number of line voltage dimmable retrofit LED housings, that are near the cost of a MR16 bulb. These housings replace the old recessed can with an integrated AC LED driver and LED bulb array.
Today, as was the case 3 years ago, commercial and residential LED recessed lights separate the enclosure from the LED driver, allowing for different color and brightness LED’s to be used, and to optimize the electronic circuitry for the type of dimmer being used. The prices are much lower compared to 3 years ago, but still high comparing a $40 halogen enclosure to a similar style $150 LED enclosure. It is possible to replace the entire enclosure, but it is a big job requiring ripping out ceiling drywall.
12V MR16’s can be powered by 12V AC electromagnetic transformers or electronic low voltage drivers. ELV drivers offer much higher efficiencies, but require compatible dimmers, and sometimes dimmers specifically designed for ELV drivers. Halogens are almost always powered by electromagnetic transformers due to the reduced cost and complexity. It is possible to replace the electromagnetic transformers in the enclosure with an ELV transformer, I’ve seen the electrician replace a blown transformer, he had to bring in the “small hand guy” from his crew and even then it took a lot of blind finger fiddling.
I use a Vantage Controls InFusion home automation lighting control system in my house. The system supports line-voltage forward phase and reverse phase dimmer modules, 0-10V control, and PWM control LED dimming. All loads in my installation are dimmed using forward phase dimmers. The recommended LED control setup is to use the 0-10V or PWM circuits, i.e. the dimming control and power lines are separate. The 0-10V / PWM control modules are about the same cost per line as AC dimmer modules, but the real cost is again in LED driver circuitry.
So what are my realistic choices:
The best quality option is to replace the halogen housings and line voltage control circuitry with native PWM control and native LED drivers. But, same as during construction, this is not a cost effective solution.
I can remove the transformers and convert the enclosure to line voltage, and use GU10 type MR16’s. But, tricky to remove the transformer, and the safety and legal state of the enclosure would be unknown after being converted to a line voltage receptacle.
I can remove the transformers and convert the enclosure to line voltage as above, but instead of using MR16 form factor bulbs, I can replace the insert with LED retrofit inserts.
I can replace the electromagnetic transformers with ELV’s to improve efficiency and dimmability. But again, a tricky job with marginal cost savings, and I still end up using 12V LED MR16’s.
I’ve opted to replace the halogen bulbs with LED’s on an as needed basis, i.e. when I need to replace a burnt trim, or blown bulb, I will replace the entire zone of lights with the same model of LED’s. My hope is that there will be ongoing improvements in product quality and performance, and ongoing reduction in costs as availability goes mainstream.
Here then is my review, more an exploration, greatly skewed by my subjective opinion vs. scientific fact, of the products I could find and test.
I initially tested the bulbs by replacing the halogens in my office, but this quickly became cumbersome, so instead I created a test bed for evaluation, trying to simulate the various dimmer and transformer types available.
I bought the the following items to match what I use in my house:
A Lutron DV-603P Diva dimmer, the kids and guest rooms have regular light switches, not automated, and use these dimmers.
Here are some pictures of the Elco enclosure, this will give you an idea of how to go about swapping the transformer, and how tight a squeeze it is:
The DV-603P is a vanilly halogen and incandescent dimmer, it works just fine with the magnetic transformer and halogen bulbs in my house, but the MR16 LED manufacturer’s compatibility guide require the use of specific low voltage magnetic or electronic low voltage dimmers. So I also bought:
I considered a more elaborate test setup, but I don’t have access to the required equipment, and the measurements would be interesting from a scientific perspective, not so much a subjective perspective. So I opted for a simpler test setup, attached to a piece of hobby board, capturing waveforms using my Rigol DS4022 scope and a Rigol RP1050D high voltage differential probe and the UltraScope software.
For transformers, I used the magnetic transformer from the Elco enclosure, and I bought three ELV’s from eBay, two from a known brand, and one unknown brand:
Advance Lite TC60W, $3 on eBay, I could not find any documentation on this product or brand.
For bulbs, I bought a variety of models from Amazon, eBay, and 1000bulbs:
Sylvania 58327: 50W Halogen MR16, 3000K, 35 Degree, 1450 CBCP.
These are the halogen bulbs I currently use, about $2.20 per bulb.
Torchstar TS010: Dimmable, 12V 4W MR16 LED, 6000K Daylight, 50 Watt Equivalent, 330 Lumen, 60 Degree Beam Angle.
I ordered a 10-pack from Amazon, the price worked out at about $5.50 per bulb. The packaging is generic, with a black marker dot indicating this to be a “pure white” variant. The bulb itself contains no markings, other than a small Torchstar sticker on the base. The bulb color is very blueish, like that of a daylight compact fluorescent bulb. I found the color to be very displeasing and distracting in my office environment, it made my color calibrated monitor screen appear yellow.
Torchstar TS010: Dimmable, 12V 4W MR16 LED, 3200K Warm White, 50 Watt Equivalent, 330 Lumen, 60 Degree Beam Angle
I ordered a 10-pack from Amazon, the price worked out to about $5.70 per bulb. Like the daylight version, the packaging is generic, with a black marker dot indicating this to be a “warm white” variant. The bulb itself contains no markings, other than a small Torchstar sticker on the base. The bulb color is pleasing, pretty close to the halogen.
Soraa Brilliant 00965: Dimmable, 12V 9W MR16 LED, 75 Watt Equivalent, 3000K, CRI 80, CBCP 1540, 590 Lumen
I ordered the bulbs from 1000bulbs, the price is about $28 per bulb. The color is pleasing but it appears to be ever so slightly bluer, more noticeable when dimmed. This bulb is bright, at 75W equivalent, almost too bright for my office as one of the bulbs is right above my head. Soraa specializes in high color quality products, and this model is from the older Brilliant Series, while I was really looking for the new Vivid Series bulbs, like the 00943, but it seems these bulbs are not yet available. I hope to find and test some when they do become available. At the price point of near $30 they are definitely specialty use, but I am interested in the supposed dimmability improvements.
Soraa Outdoor 00107: Dimmable, 12V 9.8W MR16 LED, 2700K
This is a 36W equivalent LED for outdoor use, I bought them for about $24 more than a year ago, the line has since been discontinued.
Soraa Premium 2 00249: Dimmable, 12V 11.5W MR16 LED, 3000K
I bought these more than a year ago for about $34 each, the line has since been discontinued.
Architectural LED MR16-DIM-12V: 2700K 45deg
I received samples of these MR16 LED’s from my electrician, I could not find any info on them.
eBay Dimmable CREE LED COB MR16: 6W MR16
I bought a batch of 10 warm white and 10 daylight 6W bulbs, and a 9W and a 12W. The 6W bulbs are about $3 per bulb. These bulbs worked surprisingly well and the color was good. Note that the 9W and 12W variants are longer than standard MR16’s.
I tested the transformer response by monitoring the high voltage AC input and low voltage AC output sides using the oscilloscope. I controlled the ELV transformers using the ELV dimmer and the magnetic transformer using the magnetic dimmer. I attached a 12ohm resistor for a purely resistive load, the halogen bulb, and the Torchstar LED bulb. I captured oscilloscope screenshots at full, half, and lowest dimming settings.
Here are the results for the magnetic transformer:
Here are the results for the HATCH RS12-60M-LED ELV transformer:
Here are the results for the HATCH RL12-60A ELV transformer:
Here are the results for the Advance Lite TC60W ELV transformer:
Looking at the results we can see that the response waveforms for the halogen bulb is, not surprisingly, near that of the resistor. We can see that the magnetic transformer and LED load has all sorts of inductive goodness going on. And we can see that the RL12-60W and TC60W ELV transformers are not nearly as well behaved as the RS12-60M-LED ELV that is specifically designed for LED loads.
I then proceeded to test the dimmability of the various LED bulbs, I summarize my subjective findings below:
Magnetic: Good dimming range
RS12-60M-LED: Good dimming range, slight transformer buzzing
RL12-60A: Good dimming range
TC60W: Good dimming range
Magnetic: Good dimming range, flicker at low end
RS12-60M-LED: Limited dimming range, no flicker, slight transformer buzzing
RL12-60A: Good dimming range, continuous flicker
TC60W: Good dimming range, flicker at low end
Soraa Premium 2:
Magnetic: Good dimming range, flicker at low end, very loud transformer buzzing
RS12-60M-LED: Good dimming range, flicker at low end, slight transformer buzzing
Magnetic: Good dimming range, flicker at low end, slight transformer buzzing
RS12-60M-LED: Good dimming range, slight transformer buzzing
eBay CREE COB:
Magnetic: Good dimming range, switches off before end of dim range
RS12-60M-LED: Good dimming range, slight transformer buzzing
I was surprised that the cheap $3 eBay CREE COB MR16 LED bulbs worked as well as they did. Only downside is they switch off at around 20% when using the magnetic transformer, but dim down well. I don’t know if they really contain CREE COB LED’s, but the COB array arrangement of LED’s provide an even light source.
The Torchstar bulbs have a slight flicker at the low end, but dims down all the way, a bit more expensive compared to the eBay bulbs, but US based Torchstar support may be worth the extra 1$ per bulb.
The RS12-60M-LED ELV transformer performed well with halogen and LED loads, but the buzzing sound with or without load was a disappointment. I tested with two units, both buzz. I contacted the manufacturer to find out if this is normal, or if the units I bought on eBay are faulty.
I have yet to find a MR16 LED that can be driven by a magnetic transformer that performs like halogens, my search continues.