Justin Richards
2021-03-19 01:35:33 UTC
The problem I have is the morning powered solar panel presents a 'very'
slowly rising Vin and the buck converter powered ESP8266 does not start
correctly.
Anyone know of an elegant design to provide a clean enable signal that will
deal with very very slowly rising Vin. Perhaps it has to be a belt and
braces approach with an Op-Amp type solution. The obvious simple answer is
to use a battery but there is no challenge and not what I want.
If interested, read on for a lengthy description..
I have repeatedly and successfully bench tested a design with slowly rising
Vin with 0.1V increments.
A cheap buck converter has an enable pin that was tied to Vin via 10K. I
removed and inserted 10v Zener and bias resistors. Works great on the bench
PSU but solar panels are a very different beast.
I was aware that a solar panel was a complex device to model but I was not
prepared for just how complex.
The challenge: (which I thought would be trivial)
Solar powered aquaponics water circulation system with remote monitoring
(as I work away 2 weeks) while maximising pump life (by varying duty cycle)
and eliminating the need for a battery (as they need replacing).
I built a 555 based controller that did just that. i.e turned on 10mins
every hour and clamped the supply to the pump so the pump was never over
'voltaged'. (see
https://hackaday.io/project/19863-aquaponics-solar-battery-eliminator)
However, it failed after some time while I was away due to corrosion, so
the obvious solution was to add remote monitoring in the form of an ESP8266
that sends regular http get requests to my local server. From there I can
do blah. (remotely task a family member to investigate the failure) .
Bonus of using a ESP8266 is that many other environmental conditions can be
monitored such as water temperature, did the water actually flow, water
tank levels etc.
Other problems encountered:
Voc (Solar Panel Open Circuit Voltage) is definitely not a viable way to
determine available power (I thought I could use it as a rough guide but I
was so wrong). However, load testing to produce useful data often results
in the supply rail drooping causing a ESP8266 reset at sun-up, sun-down and
cloud cover in between.
To solve i connected bulk capacitors across the ESP8266 and separating it
from the drooping supply via a diode. Which works but creates another
issue of slowing rising Vcc at start up which the ESP8266 does not like. I
already eliminated this problem with a device that only powers up the
system when the Voc exceeds 16v. A game of cat and mouse.
The current system works ok after a successful initial startup. i.e with
a two stage volt check, it performs a quick load test to determine if the
panel can deliver before it runs the pump proper. It continues to monitor
and shut off the pump if it falls below a preset value. Basically, if Voc
is over 16v proceed to 0.5s load test. If Vload > 10v, turn on pump. If
Vload < 10V halt, reset timers and try again later,
Using a 33w solar panel nominal voltage 21v Measured Voc 19v, Isc ~1.3A
midday. (manufacture date 1985)
17watt water pump 12V nominal. Work effectively for my purposes down to
about 7v.
Also observed very strange behaviour when powering the pump from a Rigol
DP831. Set to 12v 2amps, the display after startup would read ~12.5v and
the current would 'decrease' when 'increasing' the flow restriction. It
was very odd and counter intuitive . Using a scope revealed that the pump
generates periodic spikes and I had to conclude that the DP831 was
struggling to regulate given such a complex load.
It has been quite a learning experience for something that on the surface I
considered trivial.
slowly rising Vin and the buck converter powered ESP8266 does not start
correctly.
Anyone know of an elegant design to provide a clean enable signal that will
deal with very very slowly rising Vin. Perhaps it has to be a belt and
braces approach with an Op-Amp type solution. The obvious simple answer is
to use a battery but there is no challenge and not what I want.
If interested, read on for a lengthy description..
I have repeatedly and successfully bench tested a design with slowly rising
Vin with 0.1V increments.
A cheap buck converter has an enable pin that was tied to Vin via 10K. I
removed and inserted 10v Zener and bias resistors. Works great on the bench
PSU but solar panels are a very different beast.
I was aware that a solar panel was a complex device to model but I was not
prepared for just how complex.
The challenge: (which I thought would be trivial)
Solar powered aquaponics water circulation system with remote monitoring
(as I work away 2 weeks) while maximising pump life (by varying duty cycle)
and eliminating the need for a battery (as they need replacing).
I built a 555 based controller that did just that. i.e turned on 10mins
every hour and clamped the supply to the pump so the pump was never over
'voltaged'. (see
https://hackaday.io/project/19863-aquaponics-solar-battery-eliminator)
However, it failed after some time while I was away due to corrosion, so
the obvious solution was to add remote monitoring in the form of an ESP8266
that sends regular http get requests to my local server. From there I can
do blah. (remotely task a family member to investigate the failure) .
Bonus of using a ESP8266 is that many other environmental conditions can be
monitored such as water temperature, did the water actually flow, water
tank levels etc.
Other problems encountered:
Voc (Solar Panel Open Circuit Voltage) is definitely not a viable way to
determine available power (I thought I could use it as a rough guide but I
was so wrong). However, load testing to produce useful data often results
in the supply rail drooping causing a ESP8266 reset at sun-up, sun-down and
cloud cover in between.
To solve i connected bulk capacitors across the ESP8266 and separating it
from the drooping supply via a diode. Which works but creates another
issue of slowing rising Vcc at start up which the ESP8266 does not like. I
already eliminated this problem with a device that only powers up the
system when the Voc exceeds 16v. A game of cat and mouse.
The current system works ok after a successful initial startup. i.e with
a two stage volt check, it performs a quick load test to determine if the
panel can deliver before it runs the pump proper. It continues to monitor
and shut off the pump if it falls below a preset value. Basically, if Voc
is over 16v proceed to 0.5s load test. If Vload > 10v, turn on pump. If
Vload < 10V halt, reset timers and try again later,
Using a 33w solar panel nominal voltage 21v Measured Voc 19v, Isc ~1.3A
midday. (manufacture date 1985)
17watt water pump 12V nominal. Work effectively for my purposes down to
about 7v.
Also observed very strange behaviour when powering the pump from a Rigol
DP831. Set to 12v 2amps, the display after startup would read ~12.5v and
the current would 'decrease' when 'increasing' the flow restriction. It
was very odd and counter intuitive . Using a scope revealed that the pump
generates periodic spikes and I had to conclude that the DP831 was
struggling to regulate given such a complex load.
It has been quite a learning experience for something that on the surface I
considered trivial.
--
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist